diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
index 04bff26aa..1fa7704df 100644
--- a/.github/workflows/test.yml
+++ b/.github/workflows/test.yml
@@ -71,6 +71,20 @@ jobs:
             ${{ runner.OS }}-fixtures-
             ${{ runner.OS }}-
 
+      - name: Cache PeerTube pip directory
+        uses: actions/cache@v4
+        with:
+          path: |
+            ~/.cache/pip
+          key: ${{ runner.OS }}-${{ matrix.test_suite }}-pip-v1
+
+      - name: Cache Hugging Face models
+        uses: actions/cache@v4
+        with:
+          path: |
+            ~/.cache/huggingface
+          key: ${{ runner.OS }}-${{ matrix.test_suite }}-hugging-face-v1
+
       - name: Set env test variable (schedule)
         if: github.event_name != 'schedule'
         run: |
diff --git a/.gitignore b/.gitignore
index 6865442eb..cf3b25a1c 100644
--- a/.gitignore
+++ b/.gitignore
@@ -12,8 +12,11 @@ yarn-error.log
 /test4/
 /test5/
 /test6/
+
+# Big fixtures generated/downloaded on-demand
 /packages/tests/fixtures/video_high_bitrate_1080p.mp4
 /packages/tests/fixtures/video_59fps.mp4
+/packages/tests/fixtures/transcription/models-v1/
 
 # Production
 /storage
diff --git a/apps/peertube-runner/src/server/process/process.ts b/apps/peertube-runner/src/server/process/process.ts
index e8a1d7c28..b20c2297d 100644
--- a/apps/peertube-runner/src/server/process/process.ts
+++ b/apps/peertube-runner/src/server/process/process.ts
@@ -1,6 +1,7 @@
 import {
   RunnerJobLiveRTMPHLSTranscodingPayload,
   RunnerJobStudioTranscodingPayload,
+  RunnerJobTranscriptionPayload,
   RunnerJobVODAudioMergeTranscodingPayload,
   RunnerJobVODHLSTranscodingPayload,
   RunnerJobVODWebVideoTranscodingPayload
@@ -9,25 +10,41 @@ import { logger } from '../../shared/index.js'
 import { processAudioMergeTranscoding, processHLSTranscoding, ProcessOptions, processWebVideoTranscoding } from './shared/index.js'
 import { ProcessLiveRTMPHLSTranscoding } from './shared/process-live.js'
 import { processStudioTranscoding } from './shared/process-studio.js'
+import { processVideoTranscription } from './shared/process-transcription.js'
 
 export async function processJob (options: ProcessOptions) {
   const { server, job } = options
 
   logger.info(`[${server.url}] Processing job of type ${job.type}: ${job.uuid}`, { payload: job.payload })
 
-  if (job.type === 'vod-audio-merge-transcoding') {
-    await processAudioMergeTranscoding(options as ProcessOptions<RunnerJobVODAudioMergeTranscodingPayload>)
-  } else if (job.type === 'vod-web-video-transcoding') {
-    await processWebVideoTranscoding(options as ProcessOptions<RunnerJobVODWebVideoTranscodingPayload>)
-  } else if (job.type === 'vod-hls-transcoding') {
-    await processHLSTranscoding(options as ProcessOptions<RunnerJobVODHLSTranscodingPayload>)
-  } else if (job.type === 'live-rtmp-hls-transcoding') {
-    await new ProcessLiveRTMPHLSTranscoding(options as ProcessOptions<RunnerJobLiveRTMPHLSTranscodingPayload>).process()
-  } else if (job.type === 'video-studio-transcoding') {
-    await processStudioTranscoding(options as ProcessOptions<RunnerJobStudioTranscodingPayload>)
-  } else {
-    logger.error(`Unknown job ${job.type} to process`)
-    return
+  switch (job.type) {
+    case 'vod-audio-merge-transcoding':
+      await processAudioMergeTranscoding(options as ProcessOptions<RunnerJobVODAudioMergeTranscodingPayload>)
+      break
+
+    case 'vod-web-video-transcoding':
+      await processWebVideoTranscoding(options as ProcessOptions<RunnerJobVODWebVideoTranscodingPayload>)
+      break
+
+    case 'vod-hls-transcoding':
+      await processHLSTranscoding(options as ProcessOptions<RunnerJobVODHLSTranscodingPayload>)
+      break
+
+    case 'live-rtmp-hls-transcoding':
+      await new ProcessLiveRTMPHLSTranscoding(options as ProcessOptions<RunnerJobLiveRTMPHLSTranscodingPayload>).process()
+      break
+
+    case 'video-studio-transcoding':
+      await processStudioTranscoding(options as ProcessOptions<RunnerJobStudioTranscodingPayload>)
+      break
+
+    case 'video-transcription':
+      await processVideoTranscription(options as ProcessOptions<RunnerJobTranscriptionPayload>)
+      break
+
+    default:
+      logger.error(`Unknown job ${job.type} to process`)
+      return
   }
 
   logger.info(`[${server.url}] Finished processing job of type ${job.type}: ${job.uuid}`)
diff --git a/apps/peertube-runner/src/server/process/shared/common.ts b/apps/peertube-runner/src/server/process/shared/common.ts
index 09241d93b..cf7682991 100644
--- a/apps/peertube-runner/src/server/process/shared/common.ts
+++ b/apps/peertube-runner/src/server/process/shared/common.ts
@@ -5,7 +5,7 @@ import { RunnerJob, RunnerJobPayload } from '@peertube/peertube-models'
 import { buildUUID } from '@peertube/peertube-node-utils'
 import { PeerTubeServer } from '@peertube/peertube-server-commands'
 import { ConfigManager, downloadFile, logger } from '../../../shared/index.js'
-import { getTranscodingLogger } from './transcoding-logger.js'
+import { getWinstonLogger } from './winston-logger.js'
 
 export type JobWithToken <T extends RunnerJobPayload = RunnerJobPayload> = RunnerJob<T> & { jobToken: string }
 
@@ -101,6 +101,6 @@ function getCommonFFmpegOptions () {
       available: getDefaultAvailableEncoders(),
       encodersToTry: getDefaultEncodersToTry()
     },
-    logger: getTranscodingLogger()
+    logger: getWinstonLogger()
   }
 }
diff --git a/apps/peertube-runner/src/server/process/shared/index.ts b/apps/peertube-runner/src/server/process/shared/index.ts
index 638bf127f..67d556f91 100644
--- a/apps/peertube-runner/src/server/process/shared/index.ts
+++ b/apps/peertube-runner/src/server/process/shared/index.ts
@@ -1,3 +1,3 @@
 export * from './common.js'
 export * from './process-vod.js'
-export * from './transcoding-logger.js'
+export * from './winston-logger.js'
diff --git a/apps/peertube-runner/src/server/process/shared/process-transcription.ts b/apps/peertube-runner/src/server/process/shared/process-transcription.ts
new file mode 100644
index 000000000..715433f09
--- /dev/null
+++ b/apps/peertube-runner/src/server/process/shared/process-transcription.ts
@@ -0,0 +1,79 @@
+import { hasAudioStream } from '@peertube/peertube-ffmpeg'
+import { RunnerJobTranscriptionPayload, TranscriptionSuccess } from '@peertube/peertube-models'
+import { buildSUUID } from '@peertube/peertube-node-utils'
+import { TranscriptionModel, WhisperBuiltinModel, transcriberFactory } from '@peertube/peertube-transcription'
+import { remove } from 'fs-extra/esm'
+import { join } from 'path'
+import { ConfigManager } from '../../../shared/config-manager.js'
+import { logger } from '../../../shared/index.js'
+import { ProcessOptions, downloadInputFile, scheduleTranscodingProgress } from './common.js'
+import { getWinstonLogger } from './winston-logger.js'
+
+export async function processVideoTranscription (options: ProcessOptions<RunnerJobTranscriptionPayload>) {
+  const { server, job, runnerToken } = options
+
+  const config = ConfigManager.Instance.getConfig().transcription
+
+  const payload = job.payload
+
+  let inputPath: string
+
+  const updateProgressInterval = scheduleTranscodingProgress({
+    job,
+    server,
+    runnerToken,
+    progressGetter: () => undefined
+  })
+
+  const outputPath = join(ConfigManager.Instance.getTranscriptionDirectory(), buildSUUID())
+
+  const transcriber = transcriberFactory.createFromEngineName({
+    engineName: config.engine,
+    enginePath: config.enginePath,
+    logger: getWinstonLogger()
+  })
+
+  try {
+    logger.info(`Downloading input file ${payload.input.videoFileUrl} for transcription job ${job.jobToken}`)
+
+    inputPath = await downloadInputFile({ url: payload.input.videoFileUrl, runnerToken, job })
+
+    logger.info(`Downloaded input file ${payload.input.videoFileUrl} for job ${job.jobToken}. Running transcription.`)
+
+    if (await hasAudioStream(inputPath) !== true) {
+      await server.runnerJobs.error({
+        jobToken: job.jobToken,
+        jobUUID: job.uuid,
+        runnerToken,
+        message: 'This input file does not contain audio'
+      })
+
+      return
+    }
+
+    const transcriptFile = await transcriber.transcribe({
+      mediaFilePath: inputPath,
+      model: config.modelPath
+        ? await TranscriptionModel.fromPath(config.modelPath)
+        : new WhisperBuiltinModel(config.model),
+      format: 'vtt',
+      transcriptDirectory: outputPath
+    })
+
+    const successBody: TranscriptionSuccess = {
+      inputLanguage: transcriptFile.language,
+      vttFile: transcriptFile.path
+    }
+
+    await server.runnerJobs.success({
+      jobToken: job.jobToken,
+      jobUUID: job.uuid,
+      runnerToken,
+      payload: successBody
+    })
+  } finally {
+    if (inputPath) await remove(inputPath)
+    if (outputPath) await remove(outputPath)
+    if (updateProgressInterval) clearInterval(updateProgressInterval)
+  }
+}
diff --git a/apps/peertube-runner/src/server/process/shared/transcoding-logger.ts b/apps/peertube-runner/src/server/process/shared/transcoding-logger.ts
deleted file mode 100644
index d0775e13b..000000000
--- a/apps/peertube-runner/src/server/process/shared/transcoding-logger.ts
+++ /dev/null
@@ -1,19 +0,0 @@
-import { LogFn } from 'pino'
-import { logger } from '../../../shared/index.js'
-
-export function getTranscodingLogger () {
-  return {
-    info: buildWinstonLogger(logger.info.bind(logger)),
-    debug: buildWinstonLogger(logger.debug.bind(logger)),
-    warn: buildWinstonLogger(logger.warn.bind(logger)),
-    error: buildWinstonLogger(logger.error.bind(logger))
-  }
-}
-
-function buildWinstonLogger (log: LogFn) {
-  return (arg1: string, arg2?: object) => {
-    if (arg2) return log(arg2, arg1)
-
-    return log(arg1)
-  }
-}
diff --git a/apps/peertube-runner/src/server/process/shared/winston-logger.ts b/apps/peertube-runner/src/server/process/shared/winston-logger.ts
new file mode 100644
index 000000000..d2958ea6a
--- /dev/null
+++ b/apps/peertube-runner/src/server/process/shared/winston-logger.ts
@@ -0,0 +1,19 @@
+import { LogFn } from 'pino'
+import { logger } from '../../../shared/index.js'
+
+export function getWinstonLogger () {
+  return {
+    info: buildLogLevelFn(logger.info.bind(logger)),
+    debug: buildLogLevelFn(logger.debug.bind(logger)),
+    warn: buildLogLevelFn(logger.warn.bind(logger)),
+    error: buildLogLevelFn(logger.error.bind(logger))
+  }
+}
+
+function buildLogLevelFn (log: LogFn) {
+  return (arg1: string, arg2?: object) => {
+    if (arg2) return log(arg2, arg1)
+
+    return log(arg1)
+  }
+}
diff --git a/apps/peertube-runner/src/server/shared/supported-job.ts b/apps/peertube-runner/src/server/shared/supported-job.ts
index d905b5de2..38cd28140 100644
--- a/apps/peertube-runner/src/server/shared/supported-job.ts
+++ b/apps/peertube-runner/src/server/shared/supported-job.ts
@@ -1,15 +1,16 @@
 import {
   RunnerJobLiveRTMPHLSTranscodingPayload,
   RunnerJobPayload,
-  RunnerJobType,
   RunnerJobStudioTranscodingPayload,
+  RunnerJobTranscriptionPayload,
+  RunnerJobType,
   RunnerJobVODAudioMergeTranscodingPayload,
   RunnerJobVODHLSTranscodingPayload,
   RunnerJobVODWebVideoTranscodingPayload,
   VideoStudioTaskPayload
 } from '@peertube/peertube-models'
 
-const supportedMatrix = {
+const supportedMatrix: { [ id in RunnerJobType ]: (payload: RunnerJobPayload) => boolean } = {
   'vod-web-video-transcoding': (_payload: RunnerJobVODWebVideoTranscodingPayload) => {
     return true
   },
@@ -29,6 +30,9 @@ const supportedMatrix = {
     if (!Array.isArray(tasks)) return false
 
     return tasks.every(t => t && supported.has(t.name))
+  },
+  'video-transcription': (_payload: RunnerJobTranscriptionPayload) => {
+    return true
   }
 }
 
diff --git a/apps/peertube-runner/src/shared/config-manager.ts b/apps/peertube-runner/src/shared/config-manager.ts
index 84a326a16..97a70204a 100644
--- a/apps/peertube-runner/src/shared/config-manager.ts
+++ b/apps/peertube-runner/src/shared/config-manager.ts
@@ -1,4 +1,5 @@
 import { parse, stringify } from '@iarna/toml'
+import { TranscriptionEngineName, WhisperBuiltinModelName } from '@peertube/peertube-transcription'
 import envPaths from 'env-paths'
 import { ensureDir, pathExists, remove } from 'fs-extra/esm'
 import { readFile, writeFile } from 'fs/promises'
@@ -24,6 +25,13 @@ type Config = {
     runnerName: string
     runnerDescription?: string
   }[]
+
+  transcription: {
+    engine: TranscriptionEngineName
+    enginePath: string | null
+    model: WhisperBuiltinModelName
+    modelPath: string | null
+  }
 }
 
 export class ConfigManager {
@@ -37,6 +45,12 @@ export class ConfigManager {
       threads: 2,
       nice: 20
     },
+    transcription: {
+      engine: 'whisper-ctranslate2',
+      enginePath: null,
+      model: 'small',
+      modelPath: null
+    },
     registeredInstances: []
   }
 
@@ -98,6 +112,10 @@ export class ConfigManager {
     return join(paths.cache, this.id, 'transcoding')
   }
 
+  getTranscriptionDirectory () {
+    return join(paths.cache, this.id, 'transcription')
+  }
+
   getSocketDirectory () {
     return join(paths.data, this.id)
   }
diff --git a/client/src/app/+admin/config/edit-custom-config/edit-basic-configuration.component.html b/client/src/app/+admin/config/edit-custom-config/edit-basic-configuration.component.html
index 3de40666f..ea8094d5f 100644
--- a/client/src/app/+admin/config/edit-custom-config/edit-basic-configuration.component.html
+++ b/client/src/app/+admin/config/edit-custom-config/edit-basic-configuration.component.html
@@ -318,7 +318,7 @@
             >
             <ng-container ngProjectAs="description">
               <span i18n [hidden]="isImportVideosHttpEnabled()">
-                  ⛔ You need to allow  import with HTTP URL to be able to activate this feature.
+                ⛔ You need to allow  import with HTTP URL to be able to activate this feature.
               </span>
             </ng-container>
             </my-peertube-checkbox>
@@ -359,7 +359,6 @@
       </ng-container>
 
       <ng-container formGroupName="storyboards">
-
         <div class="form-group">
           <my-peertube-checkbox
             inputName="storyboardsEnabled" formControlName="enabled"
@@ -370,7 +369,35 @@
             </ng-container>
           </my-peertube-checkbox>
         </div>
+      </ng-container>
 
+      <ng-container formGroupName="videoTranscription">
+        <div class="form-group">
+          <my-peertube-checkbox
+            inputName="videoTranscriptionEnabled" formControlName="enabled"
+            i18n-labelText labelText="Enable video transcription"
+          >
+            <ng-container ngProjectAs="description">
+              <span i18n>Automatically create a subtitle file of uploaded/imported VOD videos</span>
+            </ng-container>
+
+            <ng-container ngProjectAs="extra">
+              <div class="form-group" formGroupName="remoteRunners" [ngClass]="getTranscriptionRunnerDisabledClass()">
+                <my-peertube-checkbox
+                  inputName="videoTranscriptionRemoteRunnersEnabled" formControlName="enabled"
+                  i18n-labelText labelText="Enable remote runners for transcription"
+                >
+                  <ng-container ngProjectAs="description">
+                    <span i18n>
+                      Use <a routerLink="/admin/system/runners/runners-list">remote runners</a> to process transcription tasks.
+                      Remote runners has to register on your instance first.
+                    </span>
+                  </ng-container>
+                </my-peertube-checkbox>
+              </div>
+            </ng-container>
+          </my-peertube-checkbox>
+        </div>
       </ng-container>
     </div>
   </div>
diff --git a/client/src/app/+admin/config/edit-custom-config/edit-basic-configuration.component.ts b/client/src/app/+admin/config/edit-custom-config/edit-basic-configuration.component.ts
index 9f1cbc758..98510c7c3 100644
--- a/client/src/app/+admin/config/edit-custom-config/edit-basic-configuration.component.ts
+++ b/client/src/app/+admin/config/edit-custom-config/edit-basic-configuration.component.ts
@@ -137,6 +137,18 @@ export class EditBasicConfigurationComponent implements OnInit, OnChanges {
     return { 'disabled-checkbox-extra': !this.isSearchIndexEnabled() }
   }
 
+  // ---------------------------------------------------------------------------
+
+  isTranscriptionEnabled () {
+    return this.form.value['videoTranscription']['enabled'] === true
+  }
+
+  getTranscriptionRunnerDisabledClass () {
+    return { 'disabled-checkbox-extra': !this.isTranscriptionEnabled() }
+  }
+
+  // ---------------------------------------------------------------------------
+
   isAutoFollowIndexEnabled () {
     return this.form.value['followings']['instance']['autoFollowIndex']['enabled'] === true
   }
diff --git a/client/src/app/+admin/config/edit-custom-config/edit-custom-config.component.ts b/client/src/app/+admin/config/edit-custom-config/edit-custom-config.component.ts
index 6c14b5c55..a85b17165 100644
--- a/client/src/app/+admin/config/edit-custom-config/edit-custom-config.component.ts
+++ b/client/src/app/+admin/config/edit-custom-config/edit-custom-config.component.ts
@@ -267,6 +267,12 @@ export class EditCustomConfigComponent extends FormReactive implements OnInit {
           enabled: null
         }
       },
+      videoTranscription: {
+        enabled: null,
+        remoteRunners: {
+          enabled: null
+        }
+      },
       videoFile: {
         update: {
           enabled: null
diff --git a/client/src/app/+admin/overview/videos/video-list.component.ts b/client/src/app/+admin/overview/videos/video-list.component.ts
index 3eb932322..f7d88b752 100644
--- a/client/src/app/+admin/overview/videos/video-list.component.ts
+++ b/client/src/app/+admin/overview/videos/video-list.component.ts
@@ -1,7 +1,7 @@
 import { DatePipe, NgClass, NgFor, NgIf } from '@angular/common'
 import { Component, OnInit, ViewChild } from '@angular/core'
 import { ActivatedRoute, Router, RouterLink } from '@angular/router'
-import { AuthService, ConfirmService, Notifier, RestPagination, RestTable } from '@app/core'
+import { AuthService, ConfirmService, Notifier, RestPagination, RestTable, ServerService } from '@app/core'
 import { formatICU, getAbsoluteAPIUrl } from '@app/helpers'
 import { VideoDetails } from '@app/shared/shared-main/video/video-details.model'
 import { VideoFileTokenService } from '@app/shared/shared-main/video/video-file-token.service'
@@ -30,6 +30,7 @@ import {
   VideoActionsDropdownComponent
 } from '../../../shared/shared-video-miniature/video-actions-dropdown.component'
 import { VideoAdminService } from './video-admin.service'
+import { VideoCaptionService } from '@app/shared/shared-main/video-caption/video-caption.service'
 
 @Component({
   selector: 'my-video-list',
@@ -84,7 +85,8 @@ export class VideoListComponent extends RestTable <Video> implements OnInit {
     removeFiles: true,
     transcoding: true,
     studio: true,
-    stats: true
+    stats: true,
+    generateTranscription: true
   }
 
   loading = true
@@ -100,6 +102,8 @@ export class VideoListComponent extends RestTable <Video> implements OnInit {
     private videoService: VideoService,
     private videoAdminService: VideoAdminService,
     private videoBlockService: VideoBlockService,
+    private videoCaptionService: VideoCaptionService,
+    private server: ServerService,
     private videoFileTokenService: VideoFileTokenService
   ) {
     super()
@@ -109,6 +113,10 @@ export class VideoListComponent extends RestTable <Video> implements OnInit {
     return this.auth.getUser()
   }
 
+  get serverConfig () {
+    return this.server.getHTMLConfig()
+  }
+
   ngOnInit () {
     this.initialize()
 
@@ -160,6 +168,14 @@ export class VideoListComponent extends RestTable <Video> implements OnInit {
           isDisplayed: videos => videos.every(v => v.canRemoveFiles(this.authUser)),
           iconName: 'delete'
         }
+      ],
+      [
+        {
+          label: $localize`Generate caption`,
+          handler: videos => this.generateCaption(videos),
+          isDisplayed: videos => videos.every(v => v.canGenerateTranscription(this.authUser, this.serverConfig.videoTranscription.enabled)),
+          iconName: 'video-lang'
+        }
       ]
     ]
   }
@@ -399,4 +415,15 @@ export class VideoListComponent extends RestTable <Video> implements OnInit {
         error: err => this.notifier.error(err.message)
       })
   }
+
+  private generateCaption (videos: Video[]) {
+    this.videoCaptionService.generateCaption(videos.map(v => v.id))
+      .subscribe({
+        next: () => {
+          this.notifier.success($localize`Transcription jobs created.`)
+        },
+
+        error: err => this.notifier.error(err.message)
+      })
+  }
 }
diff --git a/client/src/app/+admin/system/jobs/jobs.component.ts b/client/src/app/+admin/system/jobs/jobs.component.ts
index 2ff60f8e0..122fe274c 100644
--- a/client/src/app/+admin/system/jobs/jobs.component.ts
+++ b/client/src/app/+admin/system/jobs/jobs.component.ts
@@ -69,6 +69,7 @@ export class JobsComponent extends RestTable implements OnInit {
     'video-redundancy',
     'video-studio-edition',
     'video-transcoding',
+    'video-transcription',
     'videos-views-stats'
   ]
 
diff --git a/client/src/app/+my-account/my-account-settings/my-account-notification-preferences/my-account-notification-preferences.component.ts b/client/src/app/+my-account/my-account-settings/my-account-notification-preferences/my-account-notification-preferences.component.ts
index 69a1b0882..2a8e48e25 100644
--- a/client/src/app/+my-account/my-account-settings/my-account-notification-preferences/my-account-notification-preferences.component.ts
+++ b/client/src/app/+my-account/my-account-settings/my-account-notification-preferences/my-account-notification-preferences.component.ts
@@ -49,7 +49,8 @@ export class MyAccountNotificationPreferencesComponent implements OnInit {
       abuseStateChange: $localize`One of your abuse reports has been accepted or rejected by moderators`,
       newPeerTubeVersion: $localize`A new PeerTube version is available`,
       newPluginVersion: $localize`One of your plugin/theme has a new available version`,
-      myVideoStudioEditionFinished: $localize`Video studio edition has finished`
+      myVideoStudioEditionFinished: $localize`Video studio edition has finished`,
+      myVideoTranscriptionGenerated: $localize`The transcription of your video has been generated`
     }
     this.notificationSettingGroups = [
       {
@@ -68,7 +69,8 @@ export class MyAccountNotificationPreferencesComponent implements OnInit {
           'blacklistOnMyVideo',
           'myVideoPublished',
           'myVideoImportFinished',
-          'myVideoStudioEditionFinished'
+          'myVideoStudioEditionFinished',
+          'myVideoTranscriptionGenerated'
         ]
       },
 
diff --git a/client/src/app/+videos/+video-edit/shared/video-edit.component.html b/client/src/app/+videos/+video-edit/shared/video-edit.component.html
index 9bc160bbf..57dcd2b2e 100644
--- a/client/src/app/+videos/+video-edit/shared/video-edit.component.html
+++ b/client/src/app/+videos/+video-edit/shared/video-edit.component.html
@@ -173,12 +173,8 @@
 
       <ng-template ngbNavContent>
         <div class="captions">
-
-          <div class="captions-header">
-            <button (click)="openAddCaptionModal()" class="peertube-create-button">
-              <my-global-icon iconName="add" aria-hidden="true"></my-global-icon>
-              <ng-container i18n>Add another caption</ng-container>
-            </button>
+          <div class="alert pt-alert-primary" *ngIf="displayTranscriptionInfo && isTranscriptionEnabled()" i18n>
+            A subtitle will be automatically generated from your video.
           </div>
 
           <div class="form-group" *ngFor="let videoCaption of videoCaptions">
@@ -226,6 +222,13 @@
             No captions for now.
           </div>
 
+          <div class="mt-3 mb-3">
+            <button (click)="openAddCaptionModal()" class="peertube-create-button">
+              <my-global-icon iconName="add" aria-hidden="true"></my-global-icon>
+              <ng-container i18n>Add a caption</ng-container>
+            </button>
+          </div>
+
         </div>
       </ng-template>
     </ng-container>
diff --git a/client/src/app/+videos/+video-edit/shared/video-edit.component.scss b/client/src/app/+videos/+video-edit/shared/video-edit.component.scss
index 944d21caf..fbe9d78cf 100644
--- a/client/src/app/+videos/+video-edit/shared/video-edit.component.scss
+++ b/client/src/app/+videos/+video-edit/shared/video-edit.component.scss
@@ -24,11 +24,6 @@ my-peertube-checkbox {
   }
 }
 
-.captions-header {
-  text-align: end;
-  margin-bottom: 1rem;
-}
-
 .caption-entry {
   display: flex;
   height: 40px;
diff --git a/client/src/app/+videos/+video-edit/shared/video-edit.component.ts b/client/src/app/+videos/+video-edit/shared/video-edit.component.ts
index a6aad3951..5469b7fa8 100644
--- a/client/src/app/+videos/+video-edit/shared/video-edit.component.ts
+++ b/client/src/app/+videos/+video-edit/shared/video-edit.component.ts
@@ -1,5 +1,16 @@
 import { DatePipe, NgClass, NgFor, NgIf, NgTemplateOutlet } from '@angular/common'
-import { ChangeDetectorRef, Component, EventEmitter, Input, NgZone, OnDestroy, OnInit, Output, ViewChild } from '@angular/core'
+import {
+  ChangeDetectorRef,
+  Component,
+  EventEmitter,
+  Input,
+  NgZone,
+  OnDestroy,
+  OnInit,
+  Output,
+  ViewChild,
+  booleanAttribute
+} from '@angular/core'
 import { AbstractControl, FormArray, FormGroup, FormsModule, ReactiveFormsModule, Validators } from '@angular/forms'
 import { HooksService, PluginService, ServerService } from '@app/core'
 import { removeElementFromArray } from '@app/helpers'
@@ -63,10 +74,10 @@ import { HelpComponent } from '../../../shared/shared-main/misc/help.component'
 import { EmbedComponent } from '../../../shared/shared-main/video/embed.component'
 import { LiveDocumentationLinkComponent } from '../../../shared/shared-video-live/live-documentation-link.component'
 import { I18nPrimengCalendarService } from './i18n-primeng-calendar.service'
+import { ThumbnailManagerComponent } from './thumbnail-manager/thumbnail-manager.component'
 import { VideoCaptionAddModalComponent } from './video-caption-add-modal.component'
 import { VideoCaptionEditModalContentComponent } from './video-caption-edit-modal-content/video-caption-edit-modal-content.component'
 import { VideoEditType } from './video-edit.type'
-import { ThumbnailManagerComponent } from './thumbnail-manager/thumbnail-manager.component'
 
 type VideoLanguages = VideoConstant<string> & { group?: string }
 type PluginField = {
@@ -122,15 +133,17 @@ export class VideoEditComponent implements OnInit, OnDestroy {
   @Input() publishedVideo: VideoDetails
 
   @Input() userVideoChannels: SelectChannelItem[] = []
-  @Input() forbidScheduledPublication = true
+
+  @Input({ transform: booleanAttribute }) forbidScheduledPublication = true
+  @Input({ transform: booleanAttribute }) displayTranscriptionInfo = true
 
   @Input() videoCaptions: VideoCaptionWithPathEdit[] = []
   @Input() videoSource: VideoSource
 
   @Input() videoChapters: VideoChapter[] = []
 
-  @Input() hideWaitTranscoding = false
-  @Input() updateVideoFileEnabled = false
+  @Input({ transform: booleanAttribute }) hideWaitTranscoding = false
+  @Input({ transform: booleanAttribute }) updateVideoFileEnabled = false
 
   @Input() type: VideoEditType
   @Input() liveVideo: LiveVideo
@@ -405,6 +418,10 @@ export class VideoEditComponent implements OnInit, OnDestroy {
     return !!this.form.value['originallyPublishedAt']
   }
 
+  isTranscriptionEnabled () {
+    return this.serverConfig.videoTranscription.enabled
+  }
+
   // ---------------------------------------------------------------------------
 
   resetField (name: string) {
diff --git a/client/src/app/+videos/+video-edit/video-add-components/video-go-live.component.html b/client/src/app/+videos/+video-edit/video-add-components/video-go-live.component.html
index 4831b7680..dceae4129 100644
--- a/client/src/app/+videos/+video-edit/video-add-components/video-go-live.component.html
+++ b/client/src/app/+videos/+video-edit/video-add-components/video-go-live.component.html
@@ -53,8 +53,8 @@
 <form [hidden]="!isInUpdateForm" novalidate [formGroup]="form">
   <my-video-edit
     [form]="form" [formErrors]="formErrors" [videoCaptions]="videoCaptions"
-    [forbidScheduledPublication]="true" [hideWaitTranscoding]="true"
     [validationMessages]="validationMessages" [userVideoChannels]="userVideoChannels" [liveVideo]="liveVideo"
+    forbidScheduledPublication="true" hideWaitTranscoding="true" displayTranscriptionInfo="false"
     type="go-live"
   ></my-video-edit>
 
diff --git a/client/src/app/+videos/+video-edit/video-add-components/video-import-torrent.component.html b/client/src/app/+videos/+video-edit/video-add-components/video-import-torrent.component.html
index af0580dc8..5dc9dd013 100644
--- a/client/src/app/+videos/+video-edit/video-add-components/video-import-torrent.component.html
+++ b/client/src/app/+videos/+video-edit/video-add-components/video-import-torrent.component.html
@@ -59,8 +59,9 @@
 <!-- Hidden because we want to load the component -->
 <form [hidden]="!hasImportedVideo" novalidate [formGroup]="form">
   <my-video-edit
-    [form]="form" [formErrors]="formErrors" [videoCaptions]="videoCaptions" [forbidScheduledPublication]="true"
+    [form]="form" [formErrors]="formErrors" [videoCaptions]="videoCaptions"
     [validationMessages]="validationMessages" [userVideoChannels]="userVideoChannels"
+    forbidScheduledPublication="true"
     type="import-torrent"
   ></my-video-edit>
 
diff --git a/client/src/app/+videos/+video-edit/video-add-components/video-import-url.component.html b/client/src/app/+videos/+video-edit/video-add-components/video-import-url.component.html
index 591573d85..e7991d398 100644
--- a/client/src/app/+videos/+video-edit/video-add-components/video-import-url.component.html
+++ b/client/src/app/+videos/+video-edit/video-add-components/video-import-url.component.html
@@ -57,8 +57,9 @@
 <form [hidden]="!hasImportedVideo" novalidate [formGroup]="form">
   <my-video-edit
     #videoEdit
-    [form]="form" [formErrors]="formErrors" [videoCaptions]="videoCaptions" [forbidScheduledPublication]="true"
+    [form]="form" [formErrors]="formErrors" [videoCaptions]="videoCaptions"
     [validationMessages]="validationMessages" [userVideoChannels]="userVideoChannels"
+    forbidScheduledPublication="true"
     type="import-url"
   ></my-video-edit>
 
diff --git a/client/src/app/+videos/+video-edit/video-add-components/video-upload.component.html b/client/src/app/+videos/+video-edit/video-add-components/video-upload.component.html
index 165a3fba5..066cfb14e 100644
--- a/client/src/app/+videos/+video-edit/video-add-components/video-upload.component.html
+++ b/client/src/app/+videos/+video-edit/video-add-components/video-upload.component.html
@@ -67,7 +67,7 @@
   <my-video-edit
     [form]="form" [formErrors]="formErrors" [videoCaptions]="videoCaptions"
     [validationMessages]="validationMessages" [userVideoChannels]="userVideoChannels"
-    [forbidScheduledPublication]="false"
+    forbidScheduledPublication="false"
     type="upload"
   ></my-video-edit>
 
diff --git a/client/src/app/+videos/+video-edit/video-update.component.html b/client/src/app/+videos/+video-edit/video-update.component.html
index 0761d4c09..b04147dee 100644
--- a/client/src/app/+videos/+video-edit/video-update.component.html
+++ b/client/src/app/+videos/+video-edit/video-update.component.html
@@ -20,6 +20,7 @@
       type="update" (pluginFieldsAdded)="hydratePluginFieldsFromVideo()"
       [liveVideo]="liveVideo" [publishedVideo]="videoDetails"
       [videoSource]="videoSource" [updateVideoFileEnabled]="isUpdateVideoFileEnabled()"
+      displayTranscriptionInfo="false"
 
       (formBuilt)="onFormBuilt()"
     >
diff --git a/client/src/app/shared/shared-instance/instance-features-table.component.html b/client/src/app/shared/shared-instance/instance-features-table.component.html
index 80b6776a6..4ea71ace4 100644
--- a/client/src/app/shared/shared-instance/instance-features-table.component.html
+++ b/client/src/app/shared/shared-instance/instance-features-table.component.html
@@ -34,6 +34,13 @@
       </td>
     </tr>
 
+    <tr>
+      <th i18n class="sub-label" scope="row">Automatic transcription</th>
+      <td>
+        <my-feature-boolean [value]="serverConfig.videoTranscription.enabled"></my-feature-boolean>
+      </td>
+    </tr>
+
     <tr>
       <th i18n class="sub-label" scope="row">Video uploads</th>
       <td>
diff --git a/client/src/app/shared/shared-main/buttons/action-dropdown.component.scss b/client/src/app/shared/shared-main/buttons/action-dropdown.component.scss
index d39f7c048..03a5a86ce 100644
--- a/client/src/app/shared/shared-main/buttons/action-dropdown.component.scss
+++ b/client/src/app/shared/shared-main/buttons/action-dropdown.component.scss
@@ -63,6 +63,10 @@
 
     &.with-icon {
       @include dropdown-with-icon-item;
+
+      .icon-video-lang {
+        top: 0;
+      }
     }
 
     a,
diff --git a/client/src/app/shared/shared-main/users/user-notification.model.ts b/client/src/app/shared/shared-main/users/user-notification.model.ts
index 8e5907255..6870fb8cf 100644
--- a/client/src/app/shared/shared-main/users/user-notification.model.ts
+++ b/client/src/app/shared/shared-main/users/user-notification.model.ts
@@ -11,6 +11,7 @@ import {
   UserNotificationType,
   UserNotificationType_Type,
   UserRight,
+  VideoConstant,
   VideoInfo
 } from '@peertube/peertube-models'
 import { logger } from '@root-helpers/logger'
@@ -90,6 +91,12 @@ export class UserNotification implements UserNotificationServer {
     username: string
   }
 
+  videoCaption?: {
+    id: number
+    language: VideoConstant<string>
+    video: VideoInfo
+  }
+
   createdAt: string
   updatedAt: string
 
@@ -149,6 +156,8 @@ export class UserNotification implements UserNotificationServer {
       this.peertube = hash.peertube
       this.registration = hash.registration
 
+      this.videoCaption = hash.videoCaption
+
       this.createdAt = hash.createdAt
       this.updatedAt = hash.updatedAt
 
@@ -250,6 +259,10 @@ export class UserNotification implements UserNotificationServer {
           this.pluginQueryParams.pluginType = this.plugin.type + ''
           break
 
+        case UserNotificationType.MY_VIDEO_TRANSCRIPTION_GENERATED:
+          this.videoUrl = this.buildVideoUrl(this.videoCaption.video)
+          break
+
         case UserNotificationType.MY_VIDEO_STUDIO_EDITION_FINISHED:
           this.videoUrl = this.buildVideoUrl(this.video)
           break
diff --git a/client/src/app/shared/shared-main/video-caption/video-caption.service.ts b/client/src/app/shared/shared-main/video-caption/video-caption.service.ts
index 502b26f95..4cd3105ac 100644
--- a/client/src/app/shared/shared-main/video-caption/video-caption.service.ts
+++ b/client/src/app/shared/shared-main/video-caption/video-caption.service.ts
@@ -1,15 +1,15 @@
-import { Observable, of } from 'rxjs'
-import { catchError, map, switchMap } from 'rxjs/operators'
 import { HttpClient } from '@angular/common/http'
 import { Injectable } from '@angular/core'
 import { RestExtractor, ServerService } from '@app/core'
 import { objectToFormData } from '@app/helpers'
 import { peertubeTranslate, sortBy } from '@peertube/peertube-core-utils'
 import { ResultList, VideoCaption } from '@peertube/peertube-models'
+import { Observable, from, of } from 'rxjs'
+import { catchError, concatMap, map, switchMap, toArray } from 'rxjs/operators'
 import { environment } from '../../../../environments/environment'
-import { VideoCaptionEdit } from './video-caption-edit.model'
 import { VideoPasswordService } from '../video/video-password.service'
 import { VideoService } from '../video/video.service'
+import { VideoCaptionEdit } from './video-caption-edit.model'
 
 @Injectable()
 export class VideoCaptionService {
@@ -74,4 +74,13 @@ export class VideoCaptionService {
   getCaptionContent ({ captionPath }: Pick<VideoCaption, 'captionPath'>) {
     return this.authHttp.get(environment.originServerUrl + captionPath, { responseType: 'text' })
   }
+
+  generateCaption (videoIds: (number | string)[]) {
+    return from(videoIds)
+      .pipe(
+        concatMap(videoId => this.authHttp.post(`${VideoService.BASE_VIDEO_URL}/${videoId}/captions/generate`, {})),
+        toArray(),
+        catchError(err => this.restExtractor.handleError(err))
+      )
+  }
 }
diff --git a/client/src/app/shared/shared-main/video/video.model.ts b/client/src/app/shared/shared-main/video/video.model.ts
index cbfb58ce3..3a9572ec2 100644
--- a/client/src/app/shared/shared-main/video/video.model.ts
+++ b/client/src/app/shared/shared-main/video/video.model.ts
@@ -244,6 +244,10 @@ export class Video implements VideoServerModel {
       this.isUpdatableBy(user)
   }
 
+  canGenerateTranscription (user: AuthUser, transcriptionEnabled: boolean) {
+    return transcriptionEnabled && this.isLocal && user.hasRight(UserRight.UPDATE_ANY_VIDEO)
+  }
+
   // ---------------------------------------------------------------------------
 
   isOwner (user: AuthUser) {
diff --git a/client/src/app/shared/shared-video-miniature/video-actions-dropdown.component.ts b/client/src/app/shared/shared-video-miniature/video-actions-dropdown.component.ts
index f06c2da57..5fb84a966 100644
--- a/client/src/app/shared/shared-video-miniature/video-actions-dropdown.component.ts
+++ b/client/src/app/shared/shared-video-miniature/video-actions-dropdown.component.ts
@@ -11,6 +11,7 @@ import {
   DropdownButtonSize,
   DropdownDirection
 } from '../shared-main/buttons/action-dropdown.component'
+import { VideoCaptionService } from '../shared-main/video-caption/video-caption.service'
 import { RedundancyService } from '../shared-main/video/redundancy.service'
 import { VideoDetails } from '../shared-main/video/video-details.model'
 import { Video } from '../shared-main/video/video.model'
@@ -37,6 +38,7 @@ export type VideoActionsDisplayType = {
   transcoding?: boolean
   studio?: boolean
   stats?: boolean
+  generateTranscription?: boolean
 }
 
 @Component({
@@ -115,6 +117,7 @@ export class VideoActionsDropdownComponent implements OnChanges {
     private videoBlocklistService: VideoBlockService,
     private screenService: ScreenService,
     private videoService: VideoService,
+    private videoCaptionService: VideoCaptionService,
     private redundancyService: RedundancyService,
     private serverService: ServerService
   ) { }
@@ -206,6 +209,10 @@ export class VideoActionsDropdownComponent implements OnChanges {
     return this.video.isLiveInfoAvailableBy(this.user)
   }
 
+  canGenerateTranscription () {
+    return this.video.canGenerateTranscription(this.user, this.serverService.getHTMLConfig().videoTranscription.enabled)
+  }
+
   isVideoDownloadableByAnonymous () {
     return (
       this.video &&
@@ -338,7 +345,7 @@ export class VideoActionsDropdownComponent implements OnChanges {
     this.videoService.runTranscoding({ videos: [ video ], type, askForForceTranscodingIfNeeded: true })
       .subscribe({
         next: () => {
-          this.notifier.success($localize`Transcoding jobs created for "${video.name}".`)
+          this.notifier.success($localize`Transcoding job created for "${video.name}".`)
           this.transcodingCreated.emit()
         },
 
@@ -346,6 +353,17 @@ export class VideoActionsDropdownComponent implements OnChanges {
       })
   }
 
+  generateCaption (video: Video) {
+    this.videoCaptionService.generateCaption([ video.id ])
+      .subscribe({
+        next: () => {
+          this.notifier.success($localize`Transcription job created for "${video.name}".`)
+        },
+
+        error: err => this.notifier.error(err.message)
+      })
+  }
+
   onVideoBlocked () {
     this.videoBlocked.emit()
   }
@@ -466,6 +484,14 @@ export class VideoActionsDropdownComponent implements OnChanges {
           iconName: 'delete'
         }
       ],
+      [
+        {
+          label: $localize`Generate caption`,
+          handler: ({ video }) => this.generateCaption(video),
+          isDisplayed: () => this.displayOptions.generateTranscription && this.canGenerateTranscription(),
+          iconName: 'video-lang'
+        }
+      ],
       [ // actions regarding the account/its server
         {
           label: $localize`Mute account`,
diff --git a/client/src/app/shared/standalone-notifications/user-notifications.component.html b/client/src/app/shared/standalone-notifications/user-notifications.component.html
index d3367ede9..c0c824221 100644
--- a/client/src/app/shared/standalone-notifications/user-notifications.component.html
+++ b/client/src/app/shared/standalone-notifications/user-notifications.component.html
@@ -239,6 +239,14 @@
         }
       </ng-container>
 
+      <ng-container *ngSwitchCase="22"> <!-- UserNotificationType.MY_VIDEO_TRANSCRIPTION_GENERATED -->
+        <my-global-icon iconName="video-lang" aria-hidden="true"></my-global-icon>
+
+        <div class="message" i18n>
+          <em>{{ notification.videoCaption.language.label }}</em> transcription of <a (click)="markAsRead(notification)" [routerLink]="notification.videoUrl">your video {{ notification.videoCaption.video.name }}</a> has been generated
+        </div>
+      </ng-container>
+
       <ng-container *ngSwitchDefault>
         <my-global-icon iconName="alert" aria-hidden="true"></my-global-icon>
 
diff --git a/config/default.yaml b/config/default.yaml
index f097588ff..8920e0d37 100644
--- a/config/default.yaml
+++ b/config/default.yaml
@@ -716,6 +716,34 @@ video_studio:
   remote_runners:
     enabled: false
 
+video_transcription:
+  # Enable automatic transcription of videos
+  enabled: false
+
+  # Choose engine for local transcription
+  # Supported: 'openai-whisper' or 'whisper-ctranslate2'
+  engine: 'whisper-ctranslate2'
+
+  # You can set a custom engine path for local transcription
+  # If not provided, PeerTube will try to automatically install it in the PeerTube bin directory
+  engine_path: null
+
+  # Choose engine model for local transcription
+  # Available for 'openai-whisper' and 'whisper-ctranslate2': 'tiny', 'base', 'small', 'medium' or 'large-v3'
+  model: 'small'
+
+  # Or specify the model path:
+  #  * PyTorch model file path for 'openai-whisper'
+  #  * CTranslate2 Whisper model directory path for 'whisper-ctranslate2'
+  # If not provided, PeerTube will automatically download the model
+  model_path: null
+
+  # Enable remote runners to transcribe videos
+  # If enabled, your instance won't transcribe the videos itself
+  # At least 1 remote runner must be configured to transcribe your videos
+  remote_runners:
+    enabled: false
+
 video_file:
   update:
     # Add ability for users to replace the video file of an existing video
diff --git a/config/production.yaml.example b/config/production.yaml.example
index 0e8fd86ee..0296b7e56 100644
--- a/config/production.yaml.example
+++ b/config/production.yaml.example
@@ -726,6 +726,34 @@ video_studio:
   remote_runners:
     enabled: false
 
+video_transcription:
+  # Enable automatic transcription of videos
+  enabled: false
+
+  # Choose engine for local transcription
+  # Supported: 'openai-whisper' or 'whisper-ctranslate2'
+  engine: 'whisper-ctranslate2'
+
+  # You can set a custom engine path for local transcription
+  # If not provided, PeerTube will try to automatically install it in the PeerTube bin directory
+  engine_path: null
+
+  # Choose engine model for local transcription
+  # Available for 'openai-whisper' and 'whisper-ctranslate2': 'tiny', 'base', 'small', 'medium' or 'large-v3'
+  model: 'small'
+
+  # Or specify the model path:
+  #  * PyTorch model file path for 'openai-whisper'
+  #  * CTranslate2 Whisper model directory path for 'whisper-ctranslate2'
+  # If not provided, PeerTube will automatically download the model
+  model_path: null
+
+  # Enable remote runners to transcribe videos
+  # If enabled, your instance won't transcribe the videos itself
+  # At least 1 remote runner must be configured to transcribe your videos
+  remote_runners:
+    enabled: false
+
 video_file:
   update:
     # Add ability for users to replace the video file of an existing video
diff --git a/config/test.yaml b/config/test.yaml
index e3ac19cef..4be2d25d9 100644
--- a/config/test.yaml
+++ b/config/test.yaml
@@ -166,3 +166,6 @@ open_telemetry:
 search:
   search_index:
     url: 'https://search.joinpeertube.org/'
+
+video_transcription:
+  model: 'tiny'
diff --git a/packages/ffmpeg/src/ffmpeg-command-wrapper.ts b/packages/ffmpeg/src/ffmpeg-command-wrapper.ts
index 482d20320..9dd867299 100644
--- a/packages/ffmpeg/src/ffmpeg-command-wrapper.ts
+++ b/packages/ffmpeg/src/ffmpeg-command-wrapper.ts
@@ -1,14 +1,13 @@
 import { pick, promisify0 } from '@peertube/peertube-core-utils'
-import { AvailableEncoders, EncoderOptionsBuilder, EncoderOptionsBuilderParams, EncoderProfile } from '@peertube/peertube-models'
+import {
+  AvailableEncoders,
+  EncoderOptionsBuilder,
+  EncoderOptionsBuilderParams,
+  EncoderProfile,
+  SimpleLogger
+} from '@peertube/peertube-models'
 import ffmpeg, { FfmpegCommand } from 'fluent-ffmpeg'
 
-type FFmpegLogger = {
-  info: (msg: string, obj?: object) => void
-  debug: (msg: string, obj?: object) => void
-  warn: (msg: string, obj?: object) => void
-  error: (msg: string, obj?: object) => void
-}
-
 export interface FFmpegCommandWrapperOptions {
   availableEncoders?: AvailableEncoders
   profile?: string
@@ -17,7 +16,7 @@ export interface FFmpegCommandWrapperOptions {
   tmpDirectory: string
   threads: number
 
-  logger: FFmpegLogger
+  logger: SimpleLogger
   lTags?: { tags: string[] }
 
   updateJobProgress?: (progress?: number) => void
@@ -35,7 +34,7 @@ export class FFmpegCommandWrapper {
   private readonly tmpDirectory: string
   private readonly threads: number
 
-  private readonly logger: FFmpegLogger
+  private readonly logger: SimpleLogger
   private readonly lTags: { tags: string[] }
 
   private readonly updateJobProgress: (progress?: number) => void
diff --git a/packages/jiwer/README.md b/packages/jiwer/README.md
deleted file mode 100644
index 663c399bf..000000000
--- a/packages/jiwer/README.md
+++ /dev/null
@@ -1,37 +0,0 @@
-JiWER
-=====
-__JiWER__ CLI NodeJs wrapper.
-
-> *JiWER is a python tool for computing the word-error-rate of ASR systems.*
-> https://jitsi.github.io/jiwer/cli/
-
-__JiWER__ serves as a reference implementation to calculate errors rates between 2 text files:
-- WER (Word Error Rate)
-- CER (Character Error Rate)
-
-Build
------
-
-```sh
-npm run build
-```
-
-Usage
------
-```typescript
-const jiwerCLI = new JiwerClI('./reference.txt', './hypothesis.txt')
-
-// WER as a percentage, ex: 0.03 -> 3%
-console.log(await jiwerCLI.wer())
-
-// CER as a percentage: 0.01 -> 1%
-console.log(await jiwerCLI.cer())
-
-// Detailed comparison report
-console.log(await jiwerCLI.alignment())
-```
-
-Resources
----------
-- https://jitsi.github.io/jiwer/
-- https://github.com/rapidfuzz/RapidFuzz
diff --git a/packages/jiwer/requirements.txt b/packages/jiwer/requirements.txt
deleted file mode 100644
index 53d9fb34b..000000000
--- a/packages/jiwer/requirements.txt
+++ /dev/null
@@ -1 +0,0 @@
-jiwer==3.0.4
diff --git a/packages/jiwer/src/index.ts b/packages/jiwer/src/index.ts
deleted file mode 100644
index 586fd0843..000000000
--- a/packages/jiwer/src/index.ts
+++ /dev/null
@@ -1 +0,0 @@
-export * from './jiwer-cli.js'
diff --git a/packages/models/src/common/index.ts b/packages/models/src/common/index.ts
index 1f8ebdd94..846164d8f 100644
--- a/packages/models/src/common/index.ts
+++ b/packages/models/src/common/index.ts
@@ -1,2 +1,3 @@
 export * from './file-storage.enum.js'
 export * from './result-list.model.js'
+export * from './simple-logger.model.js'
diff --git a/packages/models/src/common/simple-logger.model.ts b/packages/models/src/common/simple-logger.model.ts
new file mode 100644
index 000000000..7561ac4b6
--- /dev/null
+++ b/packages/models/src/common/simple-logger.model.ts
@@ -0,0 +1,6 @@
+export type SimpleLogger = {
+  info: (msg: string, obj?: object) => void
+  debug: (msg: string, obj?: object) => void
+  warn: (msg: string, obj?: object) => void
+  error: (msg: string, obj?: object) => void
+}
diff --git a/packages/models/src/runners/runner-job-payload.model.ts b/packages/models/src/runners/runner-job-payload.model.ts
index 19b9d649b..605e216cf 100644
--- a/packages/models/src/runners/runner-job-payload.model.ts
+++ b/packages/models/src/runners/runner-job-payload.model.ts
@@ -8,7 +8,8 @@ export type RunnerJobVODPayload =
 export type RunnerJobPayload =
   RunnerJobVODPayload |
   RunnerJobLiveRTMPHLSTranscodingPayload |
-  RunnerJobStudioTranscodingPayload
+  RunnerJobStudioTranscodingPayload |
+  RunnerJobTranscriptionPayload
 
 // ---------------------------------------------------------------------------
 
@@ -54,6 +55,12 @@ export interface RunnerJobStudioTranscodingPayload {
   tasks: VideoStudioTaskPayload[]
 }
 
+export interface RunnerJobTranscriptionPayload {
+  input: {
+    videoFileUrl: string
+  }
+}
+
 // ---------------------------------------------------------------------------
 
 export function isAudioMergeTranscodingPayload (payload: RunnerJobPayload): payload is RunnerJobVODAudioMergeTranscodingPayload {
diff --git a/packages/models/src/runners/runner-job-private-payload.model.ts b/packages/models/src/runners/runner-job-private-payload.model.ts
index 3c9f08de6..3a1398bc7 100644
--- a/packages/models/src/runners/runner-job-private-payload.model.ts
+++ b/packages/models/src/runners/runner-job-private-payload.model.ts
@@ -8,7 +8,8 @@ export type RunnerJobVODPrivatePayload =
 export type RunnerJobPrivatePayload =
   RunnerJobVODPrivatePayload |
   RunnerJobLiveRTMPHLSTranscodingPrivatePayload |
-  RunnerJobVideoStudioTranscodingPrivatePayload
+  RunnerJobVideoStudioTranscodingPrivatePayload |
+  RunnerJobTranscriptionPrivatePayload
 
 // ---------------------------------------------------------------------------
 
@@ -45,3 +46,9 @@ export interface RunnerJobVideoStudioTranscodingPrivatePayload {
   videoUUID: string
   originalTasks: VideoStudioTaskPayload[]
 }
+
+// ---------------------------------------------------------------------------
+
+export interface RunnerJobTranscriptionPrivatePayload {
+  videoUUID: string
+}
diff --git a/packages/models/src/runners/runner-job-success-body.model.ts b/packages/models/src/runners/runner-job-success-body.model.ts
index f45336b05..541f815ca 100644
--- a/packages/models/src/runners/runner-job-success-body.model.ts
+++ b/packages/models/src/runners/runner-job-success-body.model.ts
@@ -12,7 +12,8 @@ export type RunnerJobSuccessPayload =
   VODHLSTranscodingSuccess |
   VODAudioMergeTranscodingSuccess |
   LiveRTMPHLSTranscodingSuccess |
-  VideoStudioTranscodingSuccess
+  VideoStudioTranscodingSuccess |
+  TranscriptionSuccess
 
 export interface VODWebVideoTranscodingSuccess {
   videoFile: Blob | string
@@ -35,6 +36,12 @@ export interface VideoStudioTranscodingSuccess {
   videoFile: Blob | string
 }
 
+export interface TranscriptionSuccess {
+  inputLanguage: string
+
+  vttFile: Blob | string
+}
+
 export function isWebVideoOrAudioMergeTranscodingPayloadSuccess (
   payload: RunnerJobSuccessPayload
 ): payload is VODHLSTranscodingSuccess | VODAudioMergeTranscodingSuccess {
@@ -44,3 +51,7 @@ export function isWebVideoOrAudioMergeTranscodingPayloadSuccess (
 export function isHLSTranscodingPayloadSuccess (payload: RunnerJobSuccessPayload): payload is VODHLSTranscodingSuccess {
   return !!(payload as VODHLSTranscodingSuccess)?.resolutionPlaylistFile
 }
+
+export function isTranscriptionPayloadSuccess (payload: RunnerJobSuccessPayload): payload is TranscriptionSuccess {
+  return !!(payload as TranscriptionSuccess)?.vttFile
+}
diff --git a/packages/models/src/runners/runner-job-type.type.ts b/packages/models/src/runners/runner-job-type.type.ts
index 91b92a729..0456e788c 100644
--- a/packages/models/src/runners/runner-job-type.type.ts
+++ b/packages/models/src/runners/runner-job-type.type.ts
@@ -3,4 +3,5 @@ export type RunnerJobType =
   'vod-hls-transcoding' |
   'vod-audio-merge-transcoding' |
   'live-rtmp-hls-transcoding' |
-  'video-studio-transcoding'
+  'video-studio-transcoding' |
+  'video-transcription'
diff --git a/packages/models/src/server/custom-config.model.ts b/packages/models/src/server/custom-config.model.ts
index a443525df..f053dc6ac 100644
--- a/packages/models/src/server/custom-config.model.ts
+++ b/packages/models/src/server/custom-config.model.ts
@@ -179,6 +179,14 @@ export interface CustomConfig {
     }
   }
 
+  videoTranscription: {
+    enabled: boolean
+
+    remoteRunners: {
+      enabled: boolean
+    }
+  }
+
   videoFile: {
     update: {
       enabled: boolean
diff --git a/packages/models/src/server/job.model.ts b/packages/models/src/server/job.model.ts
index 7713a9d3e..8402f6cb4 100644
--- a/packages/models/src/server/job.model.ts
+++ b/packages/models/src/server/job.model.ts
@@ -33,6 +33,7 @@ export type JobType =
   | 'generate-video-storyboard'
   | 'create-user-export'
   | 'import-user-archive'
+  | 'video-transcription'
 
 export interface Job {
   id: number | string
@@ -101,11 +102,16 @@ export interface VideoImportYoutubeDLPayload {
   type: VideoImportYoutubeDLPayloadType
   videoImportId: number
 
+  generateTranscription: boolean
+
   fileExt?: string
 }
 
 export interface VideoImportTorrentPayload {
   type: VideoImportTorrentPayloadType
+
+  generateTranscription: boolean
+
   videoImportId: number
 }
 
@@ -316,3 +322,9 @@ export interface CreateUserExportPayload {
 export interface ImportUserArchivePayload {
   userImportId: number
 }
+
+// ---------------------------------------------------------------------------
+
+export interface VideoTranscriptionPayload {
+  videoUUID: string
+}
diff --git a/packages/models/src/server/server-config.model.ts b/packages/models/src/server/server-config.model.ts
index 2e3626854..3d26d2efb 100644
--- a/packages/models/src/server/server-config.model.ts
+++ b/packages/models/src/server/server-config.model.ts
@@ -346,6 +346,10 @@ export interface ServerConfig {
   storyboards: {
     enabled: boolean
   }
+
+  videoTranscription: {
+    enabled: boolean
+  }
 }
 
 export type HTMLServerConfig = Omit<ServerConfig, 'signup'>
diff --git a/packages/models/src/server/server-error-code.enum.ts b/packages/models/src/server/server-error-code.enum.ts
index 156525c79..53811d55c 100644
--- a/packages/models/src/server/server-error-code.enum.ts
+++ b/packages/models/src/server/server-error-code.enum.ts
@@ -3,58 +3,59 @@ export const ServerErrorCode = {
    * The simplest form of payload too large: when the file size is over the
    * global file size limit
    */
-  MAX_FILE_SIZE_REACHED:'max_file_size_reached',
+  MAX_FILE_SIZE_REACHED: 'max_file_size_reached',
 
   /**
    * The payload is too large for the user quota set
    */
-  QUOTA_REACHED:'quota_reached',
+  QUOTA_REACHED: 'quota_reached',
 
   /**
    * Error yielded upon trying to access a video that is not federated, nor can
-   * be. This may be due to: remote videos on instances that are not followed by
+   * be. This may be due to:  remote videos on instances that are not followed by
    * yours, and with your instance disallowing unknown instances being accessed.
    */
-  DOES_NOT_RESPECT_FOLLOW_CONSTRAINTS:'does_not_respect_follow_constraints',
+  DOES_NOT_RESPECT_FOLLOW_CONSTRAINTS: 'does_not_respect_follow_constraints',
 
-  LIVE_NOT_ENABLED:'live_not_enabled',
-  LIVE_NOT_ALLOWING_REPLAY:'live_not_allowing_replay',
-  LIVE_CONFLICTING_PERMANENT_AND_SAVE_REPLAY:'live_conflicting_permanent_and_save_replay',
+  LIVE_NOT_ENABLED: 'live_not_enabled',
+  LIVE_NOT_ALLOWING_REPLAY: 'live_not_allowing_replay',
+  LIVE_CONFLICTING_PERMANENT_AND_SAVE_REPLAY: 'live_conflicting_permanent_and_save_replay',
   /**
-   * Pretty self-explanatory: the set maximum number of simultaneous lives was
+   * Pretty self-explanatory:  the set maximum number of simultaneous lives was
    * reached, and this error is typically there to inform the user trying to
    * broadcast one.
    */
-  MAX_INSTANCE_LIVES_LIMIT_REACHED:'max_instance_lives_limit_reached',
+  MAX_INSTANCE_LIVES_LIMIT_REACHED: 'max_instance_lives_limit_reached',
   /**
-   * Pretty self-explanatory: the set maximum number of simultaneous lives FOR
+   * Pretty self-explanatory:  the set maximum number of simultaneous lives FOR
    * THIS USER was reached, and this error is typically there to inform the user
    * trying to broadcast one.
    */
-  MAX_USER_LIVES_LIMIT_REACHED:'max_user_lives_limit_reached',
+  MAX_USER_LIVES_LIMIT_REACHED: 'max_user_lives_limit_reached',
 
   /**
    * A torrent should have at most one correct video file. Any more and we will
    * not be able to choose automatically.
    */
-  INCORRECT_FILES_IN_TORRENT:'incorrect_files_in_torrent',
+  INCORRECT_FILES_IN_TORRENT: 'incorrect_files_in_torrent',
 
-  COMMENT_NOT_ASSOCIATED_TO_VIDEO:'comment_not_associated_to_video',
+  COMMENT_NOT_ASSOCIATED_TO_VIDEO: 'comment_not_associated_to_video',
 
-  MISSING_TWO_FACTOR:'missing_two_factor',
-  INVALID_TWO_FACTOR:'invalid_two_factor',
+  MISSING_TWO_FACTOR: 'missing_two_factor',
+  INVALID_TWO_FACTOR: 'invalid_two_factor',
 
-  ACCOUNT_WAITING_FOR_APPROVAL:'account_waiting_for_approval',
-  ACCOUNT_APPROVAL_REJECTED:'account_approval_rejected',
+  ACCOUNT_WAITING_FOR_APPROVAL: 'account_waiting_for_approval',
+  ACCOUNT_APPROVAL_REJECTED: 'account_approval_rejected',
 
-  RUNNER_JOB_NOT_IN_PROCESSING_STATE:'runner_job_not_in_processing_state',
-  RUNNER_JOB_NOT_IN_PENDING_STATE:'runner_job_not_in_pending_state',
-  UNKNOWN_RUNNER_TOKEN:'unknown_runner_token',
+  RUNNER_JOB_NOT_IN_PROCESSING_STATE: 'runner_job_not_in_processing_state',
+  RUNNER_JOB_NOT_IN_PENDING_STATE: 'runner_job_not_in_pending_state',
+  UNKNOWN_RUNNER_TOKEN: 'unknown_runner_token',
 
-  VIDEO_REQUIRES_PASSWORD:'video_requires_password',
-  INCORRECT_VIDEO_PASSWORD:'incorrect_video_password',
+  VIDEO_REQUIRES_PASSWORD: 'video_requires_password',
+  INCORRECT_VIDEO_PASSWORD: 'incorrect_video_password',
 
-  VIDEO_ALREADY_BEING_TRANSCODED:'video_already_being_transcoded',
+  VIDEO_ALREADY_BEING_TRANSCODED: 'video_already_being_transcoded',
+  VIDEO_ALREADY_BEING_TRANSCRIBED: 'video_already_being_transcribed',
 
   MAX_USER_VIDEO_QUOTA_EXCEEDED_FOR_USER_EXPORT: 'max_user_video_quota_exceeded_for_user_export'
 } as const
diff --git a/packages/models/src/users/user-notification-setting.model.ts b/packages/models/src/users/user-notification-setting.model.ts
index fbd94994e..64d8cfb2d 100644
--- a/packages/models/src/users/user-notification-setting.model.ts
+++ b/packages/models/src/users/user-notification-setting.model.ts
@@ -31,4 +31,6 @@ export interface UserNotificationSetting {
   newPluginVersion: UserNotificationSettingValueType
 
   myVideoStudioEditionFinished: UserNotificationSettingValueType
+
+  myVideoTranscriptionGenerated: UserNotificationSettingValueType
 }
diff --git a/packages/models/src/users/user-notification.model.ts b/packages/models/src/users/user-notification.model.ts
index e8435f031..362c080ec 100644
--- a/packages/models/src/users/user-notification.model.ts
+++ b/packages/models/src/users/user-notification.model.ts
@@ -1,6 +1,7 @@
 import { FollowState } from '../actors/index.js'
 import { AbuseStateType } from '../moderation/index.js'
 import { PluginType_Type } from '../plugins/index.js'
+import { VideoConstant } from '../videos/video-constant.model.js'
 
 export const UserNotificationType = {
   NEW_VIDEO_FROM_SUBSCRIPTION: 1,
@@ -36,7 +37,9 @@ export const UserNotificationType = {
 
   NEW_USER_REGISTRATION_REQUEST: 20,
 
-  NEW_LIVE_FROM_SUBSCRIPTION: 21
+  NEW_LIVE_FROM_SUBSCRIPTION: 21,
+
+  MY_VIDEO_TRANSCRIPTION_GENERATED: 22
 } as const
 
 export type UserNotificationType_Type = typeof UserNotificationType[keyof typeof UserNotificationType]
@@ -138,6 +141,12 @@ export interface UserNotification {
     username: string
   }
 
+  videoCaption?: {
+    id: number
+    language: VideoConstant<string>
+    video: VideoInfo
+  }
+
   createdAt: string
   updatedAt: string
 }
diff --git a/packages/models/src/videos/caption/index.ts b/packages/models/src/videos/caption/index.ts
index a175768ce..bd48c2554 100644
--- a/packages/models/src/videos/caption/index.ts
+++ b/packages/models/src/videos/caption/index.ts
@@ -1,2 +1,3 @@
-export * from './video-caption.model.js'
+export * from './video-caption-generate.model.js'
 export * from './video-caption-update.model.js'
+export * from './video-caption.model.js'
diff --git a/packages/models/src/videos/caption/video-caption-generate.model.ts b/packages/models/src/videos/caption/video-caption-generate.model.ts
new file mode 100644
index 000000000..996e622ec
--- /dev/null
+++ b/packages/models/src/videos/caption/video-caption-generate.model.ts
@@ -0,0 +1,3 @@
+export interface VideoCaptionGenerate {
+  forceTranscription?: boolean // Default false
+}
diff --git a/packages/models/src/videos/import/video-import-create.model.ts b/packages/models/src/videos/import/video-import-create.model.ts
index 3ec0d22f3..81618631a 100644
--- a/packages/models/src/videos/import/video-import-create.model.ts
+++ b/packages/models/src/videos/import/video-import-create.model.ts
@@ -5,5 +5,8 @@ export interface VideoImportCreate extends VideoUpdate {
   magnetUri?: string
   torrentfile?: Blob
 
+  // Default is true if the feature is enabled by the instance admin
+  generateTranscription?: boolean
+
   channelId: number // Required
 }
diff --git a/packages/models/src/videos/video-create.model.ts b/packages/models/src/videos/video-create.model.ts
index 0eca38372..6fa72fde0 100644
--- a/packages/models/src/videos/video-create.model.ts
+++ b/packages/models/src/videos/video-create.model.ts
@@ -27,4 +27,7 @@ export interface VideoCreate {
 
   thumbnailfile?: Blob | string
   previewfile?: Blob | string
+
+  // Default is true if the feature is enabled by the instance admin
+  generateTranscription?: boolean
 }
diff --git a/packages/server-commands/src/runners/runner-jobs-command.ts b/packages/server-commands/src/runners/runner-jobs-command.ts
index 4e702199f..64bcc5655 100644
--- a/packages/server-commands/src/runners/runner-jobs-command.ts
+++ b/packages/server-commands/src/runners/runner-jobs-command.ts
@@ -5,9 +5,6 @@ import {
   AcceptRunnerJobResult,
   ErrorRunnerJobBody,
   HttpStatusCode,
-  isHLSTranscodingPayloadSuccess,
-  isLiveRTMPHLSTranscodingUpdatePayload,
-  isWebVideoOrAudioMergeTranscodingPayloadSuccess,
   ListRunnerJobsQuery,
   RequestRunnerJobBody,
   RequestRunnerJobResult,
@@ -22,8 +19,13 @@ import {
   RunnerJobType,
   RunnerJobUpdateBody,
   RunnerJobVODPayload,
+  TranscriptionSuccess,
   VODHLSTranscodingSuccess,
-  VODWebVideoTranscodingSuccess
+  VODWebVideoTranscodingSuccess,
+  isHLSTranscodingPayloadSuccess,
+  isLiveRTMPHLSTranscodingUpdatePayload,
+  isTranscriptionPayloadSuccess,
+  isWebVideoOrAudioMergeTranscodingPayloadSuccess
 } from '@peertube/peertube-models'
 import { unwrapBody } from '../requests/index.js'
 import { waitJobs } from '../server/jobs.js'
@@ -196,6 +198,12 @@ export class RunnerJobsCommand extends AbstractCommand {
       payloadWithoutFiles = omit(payloadWithoutFiles as VODHLSTranscodingSuccess, [ 'resolutionPlaylistFile' ])
     }
 
+    if (isTranscriptionPayloadSuccess(payload) && payload.vttFile) {
+      attaches[`payload[vttFile]`] = payload.vttFile
+
+      payloadWithoutFiles = omit(payloadWithoutFiles as TranscriptionSuccess, [ 'vttFile' ])
+    }
+
     return this.postUploadRequest({
       ...options,
 
diff --git a/packages/server-commands/src/server/config-command.ts b/packages/server-commands/src/server/config-command.ts
index 874fb3c5a..ed648818e 100644
--- a/packages/server-commands/src/server/config-command.ts
+++ b/packages/server-commands/src/server/config-command.ts
@@ -355,6 +355,29 @@ export class ConfigCommand extends AbstractCommand {
 
   // ---------------------------------------------------------------------------
 
+  enableTranscription ({ remote = false }: { remote?: boolean } = {}) {
+    return this.setTranscriptionEnabled(true, remote)
+  }
+
+  disableTranscription () {
+    return this.setTranscriptionEnabled(false, false)
+  }
+
+  private setTranscriptionEnabled (enabled: boolean, remoteEnabled: boolean) {
+    return this.updateExistingConfig({
+      newConfig: {
+        videoTranscription: {
+          enabled,
+          remoteRunners: {
+            enabled: remoteEnabled
+          }
+        }
+      }
+    })
+  }
+
+  // ---------------------------------------------------------------------------
+
   getConfig (options: OverrideCommandOptions = {}) {
     const path = '/api/v1/config'
 
diff --git a/packages/server-commands/src/shared/abstract-command.ts b/packages/server-commands/src/shared/abstract-command.ts
index 116b02847..665bf13f8 100644
--- a/packages/server-commands/src/shared/abstract-command.ts
+++ b/packages/server-commands/src/shared/abstract-command.ts
@@ -1,7 +1,10 @@
 /* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/no-floating-promises */
 
-import { isAbsolute } from 'path'
+import { HttpStatusCode, HttpStatusCodeType } from '@peertube/peertube-models'
 import { buildAbsoluteFixturePath, getFileSize } from '@peertube/peertube-node-utils'
+import { expect } from 'chai'
+import got, { Response as GotResponse } from 'got'
+import { isAbsolute } from 'path'
 import {
   makeDeleteRequest,
   makeGetRequest,
@@ -11,12 +14,9 @@ import {
   unwrapBody,
   unwrapText
 } from '../requests/requests.js'
-import { expect } from 'chai'
-import got, { Response as GotResponse } from 'got'
-import { HttpStatusCode, HttpStatusCodeType } from '@peertube/peertube-models'
 
-import type { PeerTubeServer } from '../server/server.js'
 import { createReadStream } from 'fs'
+import type { PeerTubeServer } from '../server/server.js'
 
 export interface OverrideCommandOptions {
   token?: string
diff --git a/packages/server-commands/src/videos/captions-command.ts b/packages/server-commands/src/videos/captions-command.ts
index a8336aa27..3834b4dbb 100644
--- a/packages/server-commands/src/videos/captions-command.ts
+++ b/packages/server-commands/src/videos/captions-command.ts
@@ -1,3 +1,4 @@
+import { pick } from '@peertube/peertube-core-utils'
 import { HttpStatusCode, ResultList, VideoCaption } from '@peertube/peertube-models'
 import { buildAbsoluteFixturePath } from '@peertube/peertube-node-utils'
 import { AbstractCommand, OverrideCommandOptions } from '../shared/index.js'
@@ -32,6 +33,23 @@ export class CaptionsCommand extends AbstractCommand {
     })
   }
 
+  runGenerate (options: OverrideCommandOptions & {
+    videoId: string | number
+    forceTranscription?: boolean
+  }) {
+    const { videoId } = options
+    const path = '/api/v1/videos/' + videoId + '/captions/generate'
+
+    return this.postBodyRequest({
+      ...options,
+
+      path,
+      fields: pick(options, [ 'forceTranscription' ]),
+      implicitToken: true,
+      defaultExpectedStatus: HttpStatusCode.NO_CONTENT_204
+    })
+  }
+
   list (options: OverrideCommandOptions & {
     videoId: string | number
     videoPassword?: string
diff --git a/packages/server-commands/src/videos/video-imports-command.ts b/packages/server-commands/src/videos/video-imports-command.ts
index c60d787c9..b0c22d332 100644
--- a/packages/server-commands/src/videos/video-imports-command.ts
+++ b/packages/server-commands/src/videos/video-imports-command.ts
@@ -4,12 +4,19 @@ import { AbstractCommand, OverrideCommandOptions } from '../shared/index.js'
 
 export class VideoImportsCommand extends AbstractCommand {
 
-  importVideo (options: OverrideCommandOptions & {
-    attributes: (VideoImportCreate | { torrentfile?: string, previewfile?: string, thumbnailfile?: string })
+  async importVideo (options: OverrideCommandOptions & {
+    attributes: (Partial<VideoImportCreate> | { torrentfile?: string, previewfile?: string, thumbnailfile?: string })
   }) {
     const { attributes } = options
     const path = '/api/v1/videos/imports'
 
+    let defaultChannelId = 1
+
+    try {
+      const { videoChannels } = await this.server.users.getMyInfo({ token: options.token })
+      defaultChannelId = videoChannels[0].id
+    } catch (e) { /* empty */ }
+
     let attaches: any = {}
     if (attributes.torrentfile) attaches = { torrentfile: attributes.torrentfile }
     if (attributes.thumbnailfile) attaches = { thumbnailfile: attributes.thumbnailfile }
@@ -20,7 +27,11 @@ export class VideoImportsCommand extends AbstractCommand {
 
       path,
       attaches,
-      fields: options.attributes,
+      fields: {
+        channelId: defaultChannelId,
+
+        ...options.attributes
+      },
       implicitToken: true,
       defaultExpectedStatus: HttpStatusCode.OK_200
     }))
diff --git a/packages/tests/fixtures/transcription/videos/communiquer-lors-dune-classe-transplantee.mp4 b/packages/tests/fixtures/transcription/videos/communiquer-lors-dune-classe-transplantee.mp4
index 4ef5b6ee1..4f18cb0c1 100644
Binary files a/packages/tests/fixtures/transcription/videos/communiquer-lors-dune-classe-transplantee.mp4 and b/packages/tests/fixtures/transcription/videos/communiquer-lors-dune-classe-transplantee.mp4 differ
diff --git a/packages/tests/fixtures/transcription/videos/derive_sectaire.mp4 b/packages/tests/fixtures/transcription/videos/derive_sectaire.mp4
index f6969a62b..6f4ebc7fd 100644
Binary files a/packages/tests/fixtures/transcription/videos/derive_sectaire.mp4 and b/packages/tests/fixtures/transcription/videos/derive_sectaire.mp4 differ
diff --git a/packages/tests/package.json b/packages/tests/package.json
index 02882ebc7..566712bf2 100644
--- a/packages/tests/package.json
+++ b/packages/tests/package.json
@@ -6,7 +6,8 @@
   "devDependencies": {},
   "scripts": {
     "build": "tsc",
-    "watch": "tsc -w"
+    "watch": "tsc -w",
+    "install-dependencies:transcription": "pip install -r ./requirements.txt ../transcription-devtools/requirements.txt"
   },
   "dependencies": {}
 }
diff --git a/packages/tests/requirements.txt b/packages/tests/requirements.txt
new file mode 100644
index 000000000..2efc72264
--- /dev/null
+++ b/packages/tests/requirements.txt
@@ -0,0 +1,2 @@
+whisper-ctranslate2
+openai-whisper
diff --git a/packages/tests/src/api/check-params/index.ts b/packages/tests/src/api/check-params/index.ts
index d92210d34..17fc996c9 100644
--- a/packages/tests/src/api/check-params/index.ts
+++ b/packages/tests/src/api/check-params/index.ts
@@ -42,6 +42,7 @@ import './video-source.js'
 import './video-storyboards.js'
 import './video-studio.js'
 import './video-token.js'
+import './video-transcription.js'
 import './videos-common-filters.js'
 import './videos-history.js'
 import './videos-overviews.js'
diff --git a/packages/tests/src/api/check-params/user-notifications.ts b/packages/tests/src/api/check-params/user-notifications.ts
index cf20324a1..cdbf09b62 100644
--- a/packages/tests/src/api/check-params/user-notifications.ts
+++ b/packages/tests/src/api/check-params/user-notifications.ts
@@ -171,6 +171,7 @@ describe('Test user notifications API validators', function () {
       abuseStateChange: UserNotificationSettingValue.WEB,
       newPeerTubeVersion: UserNotificationSettingValue.WEB,
       myVideoStudioEditionFinished: UserNotificationSettingValue.WEB,
+      myVideoTranscriptionGenerated: UserNotificationSettingValue.WEB,
       newPluginVersion: UserNotificationSettingValue.WEB
     }
 
diff --git a/packages/tests/src/api/check-params/video-transcription.ts b/packages/tests/src/api/check-params/video-transcription.ts
new file mode 100644
index 000000000..fc28924a5
--- /dev/null
+++ b/packages/tests/src/api/check-params/video-transcription.ts
@@ -0,0 +1,106 @@
+/* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/require-await */
+
+import { HttpStatusCode, UserRole } from '@peertube/peertube-models'
+import {
+  PeerTubeServer,
+  cleanupTests,
+  createMultipleServers,
+  doubleFollow,
+  setAccessTokensToServers,
+  waitJobs
+} from '@peertube/peertube-server-commands'
+
+describe('Test video transcription API validator', function () {
+  let servers: PeerTubeServer[]
+
+  let userToken: string
+  let anotherUserToken: string
+
+  let remoteId: string
+  let validId: string
+
+  // ---------------------------------------------------------------
+
+  before(async function () {
+    this.timeout(240000)
+
+    servers = await createMultipleServers(2)
+    await setAccessTokensToServers(servers)
+
+    await doubleFollow(servers[0], servers[1])
+
+    userToken = await servers[0].users.generateUserAndToken('user', UserRole.USER)
+    anotherUserToken = await servers[0].users.generateUserAndToken('user2', UserRole.USER)
+
+    {
+      const { uuid } = await servers[1].videos.quickUpload({ name: 'remote video' })
+      remoteId = uuid
+    }
+
+    {
+      const { uuid } = await servers[0].videos.quickUpload({ name: 'both 1', token: userToken })
+      validId = uuid
+    }
+
+    await waitJobs(servers)
+
+    await servers[0].config.enableTranscription()
+  })
+
+  it('Should not run transcription of an unknown video', async function () {
+    await servers[0].captions.runGenerate({ videoId: 404, expectedStatus: HttpStatusCode.NOT_FOUND_404 })
+  })
+
+  it('Should not run transcription of a remote video', async function () {
+    await servers[0].captions.runGenerate({ videoId: remoteId, expectedStatus: HttpStatusCode.BAD_REQUEST_400 })
+  })
+
+  it('Should not run transcription by a owner/moderator user', async function () {
+    await servers[0].captions.runGenerate({ videoId: validId, token: anotherUserToken, expectedStatus: HttpStatusCode.FORBIDDEN_403 })
+  })
+
+  it('Should not run transcription if a caption file already exists', async function () {
+    await servers[0].captions.add({
+      language: 'en',
+      videoId: validId,
+      fixture: 'subtitle-good1.vtt'
+    })
+
+    await servers[0].captions.runGenerate({ videoId: validId, expectedStatus: HttpStatusCode.BAD_REQUEST_400 })
+
+    await servers[0].captions.delete({ language: 'en', videoId: validId })
+  })
+
+  it('Should not run transcription if the instance disabled it', async function () {
+    await servers[0].config.disableTranscription()
+
+    await servers[0].captions.runGenerate({ videoId: validId, expectedStatus: HttpStatusCode.BAD_REQUEST_400 })
+
+    await servers[0].config.enableTranscription()
+  })
+
+  it('Should succeed to run transcription', async function () {
+    await servers[0].captions.runGenerate({ videoId: validId, token: userToken })
+  })
+
+  it('Should fail to run transcription twice', async function () {
+    await servers[0].captions.runGenerate({ videoId: validId, token: userToken, expectedStatus: HttpStatusCode.CONFLICT_409 })
+  })
+
+  it('Should fail to run transcription twice with a non-admin user with the forceTranscription boolean', async function () {
+    await servers[0].captions.runGenerate({
+      videoId: validId,
+      token: userToken,
+      forceTranscription: true,
+      expectedStatus: HttpStatusCode.FORBIDDEN_403
+    })
+  })
+
+  it('Should succeed to run transcription twice with the forceTranscription boolean', async function () {
+    await servers[0].captions.runGenerate({ videoId: validId, forceTranscription: true })
+  })
+
+  after(async function () {
+    await cleanupTests(servers)
+  })
+})
diff --git a/packages/tests/src/api/notifications/caption-notifications.ts b/packages/tests/src/api/notifications/caption-notifications.ts
new file mode 100644
index 000000000..4d584ec0b
--- /dev/null
+++ b/packages/tests/src/api/notifications/caption-notifications.ts
@@ -0,0 +1,81 @@
+/* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/require-await */
+
+import { UserNotification } from '@peertube/peertube-models'
+import { PeerTubeServer, cleanupTests, waitJobs } from '@peertube/peertube-server-commands'
+import { MockSmtpServer } from '@tests/shared/mock-servers/mock-email.js'
+import {
+  CheckerBaseParams,
+  checkMyVideoTranscriptionGenerated,
+  prepareNotificationsTest
+} from '@tests/shared/notifications.js'
+import { join } from 'path'
+
+describe('Test caption notifications', function () {
+  let servers: PeerTubeServer[] = []
+
+  let userNotifications: UserNotification[] = []
+  let emails: object[] = []
+  let userAccessToken: string
+
+  before(async function () {
+    this.timeout(120000)
+
+    const res = await prepareNotificationsTest(1)
+    emails = res.emails
+    userAccessToken = res.userAccessToken
+    servers = res.servers
+    userNotifications = res.userNotifications
+  })
+
+  describe('Transcription of my video generated is published', function () {
+    const language = { id: 'en', label: 'English' }
+    let baseParams: CheckerBaseParams
+
+    before(() => {
+      baseParams = {
+        server: servers[0],
+        emails,
+        socketNotifications: userNotifications,
+        token: userAccessToken
+      }
+    })
+
+    async function uploadAndWait () {
+      const { uuid } = await servers[0].videos.upload({
+        token: userAccessToken,
+        attributes: {
+          name: 'video',
+          fixture: join('transcription', 'videos', 'the_last_man_on_earth.mp4'),
+          language: undefined
+        }
+      })
+      await waitJobs(servers)
+
+      return servers[0].videos.get({ id: uuid })
+    }
+
+    it('Should not send a notification if transcription is not enabled', async function () {
+      this.timeout(50000)
+
+      const { name, shortUUID } = await uploadAndWait()
+
+      await checkMyVideoTranscriptionGenerated({ ...baseParams, videoName: name, shortUUID, language, checkType: 'absence' })
+    })
+
+    it('Should send a notification transcription is enabled', async function () {
+      this.timeout(240000)
+
+      await servers[0].config.enableTranscription()
+
+      const { name, shortUUID } = await uploadAndWait()
+
+      await checkMyVideoTranscriptionGenerated({ ...baseParams, videoName: name, shortUUID, language, checkType: 'presence' })
+    })
+  })
+
+  after(async function () {
+    MockSmtpServer.Instance.kill()
+
+    await cleanupTests(servers)
+  })
+})
diff --git a/packages/tests/src/api/notifications/index.ts b/packages/tests/src/api/notifications/index.ts
index d63d94182..977c4159f 100644
--- a/packages/tests/src/api/notifications/index.ts
+++ b/packages/tests/src/api/notifications/index.ts
@@ -1,4 +1,5 @@
 import './admin-notifications.js'
+import './captions-notifications.js'
 import './comments-notifications.js'
 import './moderation-notifications.js'
 import './notifications-api.js'
diff --git a/packages/tests/src/api/runners/index.ts b/packages/tests/src/api/runners/index.ts
index 441ddc874..ed8222892 100644
--- a/packages/tests/src/api/runners/index.ts
+++ b/packages/tests/src/api/runners/index.ts
@@ -2,4 +2,5 @@ export * from './runner-common.js'
 export * from './runner-live-transcoding.js'
 export * from './runner-socket.js'
 export * from './runner-studio-transcoding.js'
+export * from './runner-transcription.js'
 export * from './runner-vod-transcoding.js'
diff --git a/packages/tests/src/api/runners/runner-transcription.ts b/packages/tests/src/api/runners/runner-transcription.ts
new file mode 100644
index 000000000..cbd4b171f
--- /dev/null
+++ b/packages/tests/src/api/runners/runner-transcription.ts
@@ -0,0 +1,109 @@
+/* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/require-await */
+
+import {
+  RunnerJobTranscriptionPayload,
+  TranscriptionSuccess
+} from '@peertube/peertube-models'
+import {
+  PeerTubeServer,
+  cleanupTests,
+  createMultipleServers,
+  doubleFollow,
+  setAccessTokensToServers,
+  setDefaultVideoChannel,
+  waitJobs
+} from '@peertube/peertube-server-commands'
+import { checkPersistentTmpIsEmpty } from '@tests/shared/directories.js'
+import { expect } from 'chai'
+
+describe('Test runner transcription', function () {
+  let servers: PeerTubeServer[] = []
+  let runnerToken: string
+
+  before(async function () {
+    this.timeout(120_000)
+
+    servers = await createMultipleServers(2)
+
+    await setAccessTokensToServers(servers)
+    await setDefaultVideoChannel(servers)
+
+    await doubleFollow(servers[0], servers[1])
+
+    await servers[0].config.enableTranscription({ remote: true })
+    runnerToken = await servers[0].runners.autoRegisterRunner()
+  })
+
+  async function upload () {
+    const { uuid } = await servers[0].videos.upload({ attributes: { name: 'video', language: undefined } })
+
+    const { availableJobs } = await servers[0].runnerJobs.request({ runnerToken })
+    expect(availableJobs).to.have.lengthOf(1)
+
+    const jobUUID = availableJobs[0].uuid
+
+    const { job } = await servers[0].runnerJobs.accept<RunnerJobTranscriptionPayload>({ runnerToken, jobUUID })
+    return { uuid, job }
+  }
+
+  it('Should execute a remote transcription job', async function () {
+    this.timeout(240_000)
+
+    const { uuid, job } = await upload()
+
+    expect(job.type === 'video-transcription')
+    expect(job.payload.input.videoFileUrl).to.exist
+
+    // Check video input file
+    {
+      await servers[0].runnerJobs.getJobFile({ url: job.payload.input.videoFileUrl, jobToken: job.jobToken, runnerToken })
+    }
+
+    const payload: TranscriptionSuccess = {
+      inputLanguage: 'ar',
+      vttFile: 'subtitle-good1.vtt'
+    }
+
+    await servers[0].runnerJobs.success({ runnerToken, jobUUID: job.uuid, jobToken: job.jobToken, payload })
+
+    await waitJobs(servers)
+
+    for (const server of servers) {
+      const video = await server.videos.get({ id: uuid })
+      expect(video.language.id).to.equal('ar')
+
+      const captions = await server.captions.list({ videoId: uuid })
+      expect(captions)
+    }
+
+    await checkPersistentTmpIsEmpty(servers[0])
+  })
+
+  it('Should not assign caption/language with an unknown inputLanguage', async function () {
+    this.timeout(240_000)
+
+    const { uuid, job } = await upload()
+
+    const payload: TranscriptionSuccess = {
+      inputLanguage: 'toto',
+      vttFile: 'subtitle-good1.vtt'
+    }
+
+    await servers[0].runnerJobs.success({ runnerToken, jobUUID: job.uuid, jobToken: job.jobToken, payload })
+
+    await waitJobs(servers)
+
+    for (const server of servers) {
+      const video = await server.videos.get({ id: uuid })
+      expect(video.language.id).to.be.null
+
+      const { total, data } = await server.captions.list({ videoId: uuid })
+      expect(total).to.equal(0)
+      expect(data).to.have.lengthOf(0)
+    }
+  })
+
+  after(async function () {
+    await cleanupTests(servers)
+  })
+})
diff --git a/packages/tests/src/api/server/config.ts b/packages/tests/src/api/server/config.ts
index 179dc4b88..78728893b 100644
--- a/packages/tests/src/api/server/config.ts
+++ b/packages/tests/src/api/server/config.ts
@@ -108,6 +108,9 @@ function checkInitialConfig (server: PeerTubeServer, data: CustomConfig) {
   expect(data.videoStudio.enabled).to.be.false
   expect(data.videoStudio.remoteRunners.enabled).to.be.false
 
+  expect(data.videoTranscription.enabled).to.be.false
+  expect(data.videoTranscription.remoteRunners.enabled).to.be.false
+
   expect(data.videoFile.update.enabled).to.be.false
 
   expect(data.import.videos.concurrency).to.equal(2)
@@ -292,6 +295,12 @@ function buildNewCustomConfig (server: PeerTubeServer): CustomConfig {
         enabled: true
       }
     },
+    videoTranscription: {
+      enabled: true,
+      remoteRunners: {
+        enabled: true
+      }
+    },
     videoFile: {
       update: {
         enabled: true
diff --git a/packages/tests/src/api/videos/index.ts b/packages/tests/src/api/videos/index.ts
index a4bcd9741..358fbfaff 100644
--- a/packages/tests/src/api/videos/index.ts
+++ b/packages/tests/src/api/videos/index.ts
@@ -1,24 +1,25 @@
+import './channel-import-videos.js'
 import './multiple-servers.js'
 import './resumable-upload.js'
 import './single-server.js'
 import './video-captions.js'
 import './video-change-ownership.js'
+import './video-channel-syncs.js'
 import './video-channels.js'
 import './video-chapters.js'
-import './channel-import-videos.js'
-import './video-channel-syncs.js'
 import './video-comments.js'
 import './video-description.js'
 import './video-files.js'
 import './video-imports.js'
 import './video-nsfw.js'
-import './video-playlists.js'
 import './video-playlist-thumbnails.js'
-import './video-source.js'
+import './video-playlists.js'
 import './video-privacy.js'
 import './video-schedule-update.js'
+import './video-source.js'
+import './video-static-file-privacy.js'
+import './video-storyboard.js'
+import './video-transcription.js'
 import './videos-common-filters.js'
 import './videos-history.js'
 import './videos-overview.js'
-import './video-static-file-privacy.js'
-import './video-storyboard.js'
diff --git a/packages/tests/src/api/videos/video-transcription.ts b/packages/tests/src/api/videos/video-transcription.ts
new file mode 100644
index 000000000..6548cd9ef
--- /dev/null
+++ b/packages/tests/src/api/videos/video-transcription.ts
@@ -0,0 +1,145 @@
+/* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/require-await */
+
+import { VideoPrivacy } from '@peertube/peertube-models'
+import {
+  PeerTubeServer,
+  cleanupTests,
+  createMultipleServers,
+  doubleFollow,
+  sendRTMPStream,
+  setAccessTokensToServers,
+  setDefaultVideoChannel,
+  stopFfmpeg,
+  waitJobs
+} from '@peertube/peertube-server-commands'
+import { FIXTURE_URLS } from '@tests/shared/fixture-urls.js'
+import { checkCaption, checkLanguage, checkNoCaption, uploadForTranscription } from '@tests/shared/transcription.js'
+
+describe('Test video transcription', function () {
+  let servers: PeerTubeServer[]
+
+  before(async function () {
+    this.timeout(60000)
+
+    servers = await createMultipleServers(2)
+
+    await setAccessTokensToServers(servers)
+    await setDefaultVideoChannel(servers)
+    await doubleFollow(servers[0], servers[1])
+
+    await waitJobs(servers)
+    await waitJobs(servers)
+  })
+
+  // ---------------------------------------------------------------------------
+
+  it('Should generate a transcription on request', async function () {
+    this.timeout(360000)
+
+    await servers[0].config.disableTranscription()
+
+    const uuid = await uploadForTranscription(servers[0])
+    await waitJobs(servers)
+    await checkLanguage(servers, uuid, null)
+
+    await servers[0].config.enableTranscription()
+
+    await servers[0].captions.runGenerate({ videoId: uuid })
+    await waitJobs(servers)
+    await checkLanguage(servers, uuid, 'en')
+
+    await checkCaption(servers, uuid)
+  })
+
+  it('Should run transcription on upload by default', async function () {
+    this.timeout(360000)
+
+    const uuid = await uploadForTranscription(servers[0])
+
+    await waitJobs(servers)
+    await checkCaption(servers, uuid)
+    await checkLanguage(servers, uuid, 'en')
+  })
+
+  it('Should run transcription on import by default', async function () {
+    this.timeout(360000)
+
+    const { video } = await servers[0].videoImports.importVideo({
+      attributes: {
+        privacy: VideoPrivacy.PUBLIC,
+        targetUrl: FIXTURE_URLS.transcriptionVideo,
+        language: undefined
+      }
+    })
+
+    await waitJobs(servers)
+    await checkCaption(servers, video.uuid)
+    await checkLanguage(servers, video.uuid, 'en')
+  })
+
+  it('Should run transcription when live ended', async function () {
+    this.timeout(360000)
+
+    await servers[0].config.enableMinimumTranscoding()
+    await servers[0].config.enableLive({ allowReplay: true, transcoding: true, resolutions: 'min' })
+
+    const { live, video } = await servers[0].live.quickCreate({
+      saveReplay: true,
+      permanentLive: false,
+      privacy: VideoPrivacy.PUBLIC
+    })
+
+    const ffmpegCommand = sendRTMPStream({ rtmpBaseUrl: live.rtmpUrl, streamKey: live.streamKey })
+    await servers[0].live.waitUntilPublished({ videoId: video.id })
+
+    await stopFfmpeg(ffmpegCommand)
+
+    await servers[0].live.waitUntilReplacedByReplay({ videoId: video.id })
+    await waitJobs(servers)
+    await checkCaption(servers, video.uuid, 'WEBVTT\n\n00:')
+    await checkLanguage(servers, video.uuid, 'en')
+
+    await servers[0].config.enableLive({ allowReplay: false })
+    await servers[0].config.disableTranscoding()
+  })
+
+  it('Should not run transcription if disabled by user', async function () {
+    this.timeout(120000)
+
+    {
+      const uuid = await uploadForTranscription(servers[0], { generateTranscription: false })
+
+      await waitJobs(servers)
+      await checkNoCaption(servers, uuid)
+      await checkLanguage(servers, uuid, null)
+    }
+
+    {
+      const { video } = await servers[0].videoImports.importVideo({
+        attributes: {
+          privacy: VideoPrivacy.PUBLIC,
+          targetUrl: FIXTURE_URLS.transcriptionVideo,
+          generateTranscription: false
+        }
+      })
+
+      await waitJobs(servers)
+      await checkNoCaption(servers, video.uuid)
+      await checkLanguage(servers, video.uuid, null)
+    }
+  })
+
+  it('Should not run a transcription if the video does not contain audio', async function () {
+    this.timeout(120000)
+
+    const uuid = await uploadForTranscription(servers[0], { generateTranscription: false })
+
+    await waitJobs(servers)
+    await checkNoCaption(servers, uuid)
+    await checkLanguage(servers, uuid, null)
+  })
+
+  after(async function () {
+    await cleanupTests(servers)
+  })
+})
diff --git a/packages/tests/src/jiwer/jiwer-cli.spec.ts b/packages/tests/src/jiwer/jiwer-cli.spec.ts
index ccc2ab26b..06645f35a 100644
--- a/packages/tests/src/jiwer/jiwer-cli.spec.ts
+++ b/packages/tests/src/jiwer/jiwer-cli.spec.ts
@@ -3,7 +3,7 @@ import { buildAbsoluteFixturePath } from '@peertube/peertube-node-utils'
 import { join } from 'path'
 import { mkdir, rm, writeFile } from 'node:fs/promises'
 import { expect } from 'chai'
-import { JiwerClI } from '@peertube/peertube-jiwer'
+import { JiwerClI } from '@peertube/peertube-transcription-devtools'
 
 describe('Jiwer CLI', function () {
   const transcriptDirectory = buildAbsoluteFixturePath('transcription/transcript-evaluator')
diff --git a/packages/tests/src/peertube-runner/index.ts b/packages/tests/src/peertube-runner/index.ts
index f02611046..7807f8479 100644
--- a/packages/tests/src/peertube-runner/index.ts
+++ b/packages/tests/src/peertube-runner/index.ts
@@ -2,4 +2,5 @@ export * from './client-cli.js'
 export * from './live-transcoding.js'
 export * from './replace-file.js'
 export * from './studio-transcoding.js'
+export * from './video-transcription.js'
 export * from './vod-transcoding.js'
diff --git a/packages/tests/src/peertube-runner/live-transcoding.ts b/packages/tests/src/peertube-runner/live-transcoding.ts
index 85eac4ce9..69d44b531 100644
--- a/packages/tests/src/peertube-runner/live-transcoding.ts
+++ b/packages/tests/src/peertube-runner/live-transcoding.ts
@@ -194,7 +194,7 @@ describe('Test Live transcoding in peertube-runner program', function () {
   describe('Check cleanup', function () {
 
     it('Should have an empty cache directory', async function () {
-      await checkPeerTubeRunnerCacheIsEmpty(peertubeRunner)
+      await checkPeerTubeRunnerCacheIsEmpty(peertubeRunner, 'transcoding')
     })
   })
 
diff --git a/packages/tests/src/peertube-runner/studio-transcoding.ts b/packages/tests/src/peertube-runner/studio-transcoding.ts
index e044b2fcf..6ce390f10 100644
--- a/packages/tests/src/peertube-runner/studio-transcoding.ts
+++ b/packages/tests/src/peertube-runner/studio-transcoding.ts
@@ -111,7 +111,7 @@ describe('Test studio transcoding in peertube-runner program', function () {
   describe('Check cleanup', function () {
 
     it('Should have an empty cache directory', async function () {
-      await checkPeerTubeRunnerCacheIsEmpty(peertubeRunner)
+      await checkPeerTubeRunnerCacheIsEmpty(peertubeRunner, 'transcoding')
     })
   })
 
diff --git a/packages/tests/src/peertube-runner/video-transcription.ts b/packages/tests/src/peertube-runner/video-transcription.ts
new file mode 100644
index 000000000..a810524f5
--- /dev/null
+++ b/packages/tests/src/peertube-runner/video-transcription.ts
@@ -0,0 +1,89 @@
+/* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/require-await */
+
+import { wait } from '@peertube/peertube-core-utils'
+import { RunnerJobState } from '@peertube/peertube-models'
+import {
+  PeerTubeServer,
+  cleanupTests,
+  createMultipleServers,
+  doubleFollow,
+  setAccessTokensToServers,
+  setDefaultVideoChannel,
+  waitJobs
+} from '@peertube/peertube-server-commands'
+import { checkPeerTubeRunnerCacheIsEmpty } from '@tests/shared/directories.js'
+import { PeerTubeRunnerProcess } from '@tests/shared/peertube-runner-process.js'
+import { checkCaption, checkLanguage, checkNoCaption, uploadForTranscription } from '@tests/shared/transcription.js'
+
+describe('Test transcription in peertube-runner program', function () {
+  let servers: PeerTubeServer[] = []
+  let peertubeRunner: PeerTubeRunnerProcess
+
+  before(async function () {
+    this.timeout(120_000)
+
+    servers = await createMultipleServers(2)
+
+    await setAccessTokensToServers(servers)
+    await setDefaultVideoChannel(servers)
+
+    await doubleFollow(servers[0], servers[1])
+
+    await servers[0].config.enableTranscription({ remote: true })
+
+    const registrationToken = await servers[0].runnerRegistrationTokens.getFirstRegistrationToken()
+
+    peertubeRunner = new PeerTubeRunnerProcess(servers[0])
+    await peertubeRunner.runServer()
+    await peertubeRunner.registerPeerTubeInstance({ registrationToken, runnerName: 'runner' })
+  })
+
+  describe('Running transcription', function () {
+
+    it('Should run transcription on classic file', async function () {
+      this.timeout(360000)
+
+      const uuid = await uploadForTranscription(servers[0])
+      await waitJobs(servers, { runnerJobs: true })
+
+      await checkCaption(servers, uuid)
+      await checkLanguage(servers, uuid, 'en')
+    })
+
+    it('Should not run transcription on video without audio stream', async function () {
+      this.timeout(360000)
+
+      const uuid = await uploadForTranscription(servers[0], { fixture: 'video_short_no_audio.mp4' })
+
+      await waitJobs(servers)
+
+      let continueWhile = true
+      while (continueWhile) {
+        await wait(500)
+
+        const { data } = await servers[0].runnerJobs.list({ stateOneOf: [ RunnerJobState.ERRORED ] })
+
+        continueWhile = !data.some(j => j.type === 'video-transcription')
+      }
+
+      await checkNoCaption(servers, uuid)
+      await checkLanguage(servers, uuid, null)
+    })
+  })
+
+  describe('Check cleanup', function () {
+
+    it('Should have an empty cache directory', async function () {
+      await checkPeerTubeRunnerCacheIsEmpty(peertubeRunner, 'transcription')
+    })
+  })
+
+  after(async function () {
+    if (peertubeRunner) {
+      await peertubeRunner.unregisterPeerTubeInstance({ runnerName: 'runner' })
+      peertubeRunner.kill()
+    }
+
+    await cleanupTests(servers)
+  })
+})
diff --git a/packages/tests/src/peertube-runner/vod-transcoding.ts b/packages/tests/src/peertube-runner/vod-transcoding.ts
index da0f26079..a7397820a 100644
--- a/packages/tests/src/peertube-runner/vod-transcoding.ts
+++ b/packages/tests/src/peertube-runner/vod-transcoding.ts
@@ -373,7 +373,7 @@ describe('Test VOD transcoding in peertube-runner program', function () {
   describe('Check cleanup', function () {
 
     it('Should have an empty cache directory', async function () {
-      await checkPeerTubeRunnerCacheIsEmpty(peertubeRunner)
+      await checkPeerTubeRunnerCacheIsEmpty(peertubeRunner, 'transcoding')
     })
   })
 
diff --git a/packages/tests/src/shared/directories.ts b/packages/tests/src/shared/directories.ts
index f21e7b7c6..e9be0c77d 100644
--- a/packages/tests/src/shared/directories.ts
+++ b/packages/tests/src/shared/directories.ts
@@ -32,13 +32,13 @@ export async function checkDirectoryIsEmpty (server: PeerTubeServer, directory:
   expect(filtered).to.have.lengthOf(0)
 }
 
-export async function checkPeerTubeRunnerCacheIsEmpty (runner: PeerTubeRunnerProcess) {
-  const directoryPath = join(homedir(), '.cache', 'peertube-runner-nodejs', runner.getId(), 'transcoding')
+export async function checkPeerTubeRunnerCacheIsEmpty (runner: PeerTubeRunnerProcess, subDir: 'transcoding' | 'transcription') {
+  const directoryPath = join(homedir(), '.cache', 'peertube-runner-nodejs', runner.getId(), subDir)
 
   const directoryExists = await pathExists(directoryPath)
   expect(directoryExists).to.be.true
 
   const files = await readdir(directoryPath)
 
-  expect(files, 'Directory content: ' + files.join(', ')).to.have.lengthOf(0)
+  expect(files, `Sub-directory ${subDir} content: ${files.join(', ')}`).to.have.lengthOf(0)
 }
diff --git a/packages/tests/src/shared/fixture-urls.ts b/packages/tests/src/shared/fixture-urls.ts
index b4c659e76..daade2d83 100644
--- a/packages/tests/src/shared/fixture-urls.ts
+++ b/packages/tests/src/shared/fixture-urls.ts
@@ -26,6 +26,7 @@ export const FIXTURE_URLS = {
   badVideo: 'https://download.cpy.re/peertube/bad_video.mp4',
   goodVideo: 'https://download.cpy.re/peertube/good_video.mp4',
   goodVideo720: 'https://download.cpy.re/peertube/good_video_720.mp4',
+  transcriptionVideo: 'https://download.cpy.re/peertube/the_last_man_on_earth.mp4',
 
   chatersVideo: 'https://download.cpy.re/peertube/video_chapters.mp4',
 
diff --git a/packages/tests/src/shared/notifications.ts b/packages/tests/src/shared/notifications.ts
index 5f4b240f9..eb9ef3e35 100644
--- a/packages/tests/src/shared/notifications.ts
+++ b/packages/tests/src/shared/notifications.ts
@@ -54,6 +54,7 @@ function getAllNotificationsSettings (): UserNotificationSetting {
     autoInstanceFollowing: UserNotificationSettingValue.WEB | UserNotificationSettingValue.EMAIL,
     newPeerTubeVersion: UserNotificationSettingValue.WEB | UserNotificationSettingValue.EMAIL,
     myVideoStudioEditionFinished: UserNotificationSettingValue.WEB | UserNotificationSettingValue.EMAIL,
+    myVideoTranscriptionGenerated: UserNotificationSettingValue.WEB | UserNotificationSettingValue.EMAIL,
     newPluginVersion: UserNotificationSettingValue.WEB | UserNotificationSettingValue.EMAIL
   }
 }
@@ -758,6 +759,40 @@ async function checkNewPluginVersion (options: CheckerBaseParams & {
   await checkNotification({ ...options, notificationChecker, emailNotificationFinder })
 }
 
+async function checkMyVideoTranscriptionGenerated (options: CheckerBaseParams & {
+  videoName: string
+  shortUUID: string
+  language: {
+    id: string
+    label: string
+  }
+  checkType: CheckerType
+}) {
+  const { videoName, shortUUID, language } = options
+  const notificationType = UserNotificationType.MY_VIDEO_TRANSCRIPTION_GENERATED
+
+  function notificationChecker (notification: UserNotification, checkType: CheckerType) {
+    if (checkType === 'presence') {
+      expect(notification).to.not.be.undefined
+      expect(notification.type).to.equal(notificationType)
+
+      expect(notification.videoCaption).to.exist
+      expect(notification.videoCaption.language.id).to.equal(language.id)
+      expect(notification.videoCaption.language.label).to.equal(language.label)
+      checkVideo(notification.videoCaption.video, videoName, shortUUID)
+    } else {
+      expect(notification.videoCaption).to.satisfy(c => c === undefined || c.Video.shortUUID !== shortUUID)
+    }
+  }
+
+  function emailNotificationFinder (email: object) {
+    const text: string = email['text']
+    return text.includes(shortUUID) && text.includes('Transcription in ' + language.label)
+  }
+
+  await checkNotification({ ...options, notificationChecker, emailNotificationFinder })
+}
+
 async function prepareNotificationsTest (serversCount = 3, overrideConfigArg: any = {}) {
   const userNotifications: UserNotification[] = []
   const adminNotifications: UserNotification[] = []
@@ -863,7 +898,8 @@ export {
   checkNewPeerTubeVersion,
   checkNewPluginVersion,
   checkVideoStudioEditionIsFinished,
-  checkRegistrationRequest
+  checkRegistrationRequest,
+  checkMyVideoTranscriptionGenerated
 }
 
 // ---------------------------------------------------------------------------
diff --git a/packages/tests/src/shared/transcription.ts b/packages/tests/src/shared/transcription.ts
new file mode 100644
index 000000000..9deb153f0
--- /dev/null
+++ b/packages/tests/src/shared/transcription.ts
@@ -0,0 +1,81 @@
+/* eslint-disable @typescript-eslint/no-unused-expressions,@typescript-eslint/require-await */
+
+import { buildAbsoluteFixturePath } from '@peertube/peertube-node-utils'
+import { PeerTubeServer, VideoEdit } from '@peertube/peertube-server-commands'
+import { downloadFile, unzip } from '@peertube/peertube-transcription-devtools'
+import { expect } from 'chai'
+import { ensureDir, pathExists } from 'fs-extra/esm'
+import { join } from 'path'
+import { testCaptionFile } from './captions.js'
+import { FIXTURE_URLS } from './fixture-urls.js'
+
+type CustomModelName = 'tiny.pt' | 'faster-whisper-tiny'
+
+export async function downloadCustomModelsIfNeeded (modelName: CustomModelName) {
+  if (await pathExists(getCustomModelPath(modelName))) return
+
+  await ensureDir(getCustomModelDirectory())
+  await unzip(await downloadFile(FIXTURE_URLS.transcriptionModels, getCustomModelDirectory()))
+}
+
+export function getCustomModelDirectory () {
+  return buildAbsoluteFixturePath(join('transcription', 'models-v1'))
+}
+
+export function getCustomModelPath (modelName: CustomModelName) {
+  return join(getCustomModelDirectory(), 'models', modelName)
+}
+
+// ---------------------------------------------------------------------------
+
+export async function checkCaption (servers: PeerTubeServer[], uuid: string, captionContains = 'WEBVTT\n\n00:00.000 --> 00:') {
+  for (const server of servers) {
+    const body = await server.captions.list({ videoId: uuid })
+    expect(body.total).to.equal(1)
+    expect(body.data).to.have.lengthOf(1)
+
+    const caption = body.data[0]
+    expect(caption.language.id).to.equal('en')
+    expect(caption.language.label).to.equal('English')
+
+    {
+      await testCaptionFile(server.url, caption.captionPath, captionContains)
+    }
+  }
+}
+
+export async function checkNoCaption (servers: PeerTubeServer[], uuid: string) {
+  for (const server of servers) {
+    const body = await server.captions.list({ videoId: uuid })
+    expect(body.total).to.equal(0)
+    expect(body.data).to.have.lengthOf(0)
+  }
+}
+
+// ---------------------------------------------------------------------------
+
+export async function checkLanguage (servers: PeerTubeServer[], uuid: string, expected: string | null) {
+  for (const server of servers) {
+    const video = await server.videos.get({ id: uuid })
+
+    if (expected) {
+      expect(video.language.id).to.equal(expected)
+    } else {
+      expect(video.language.id).to.be.null
+    }
+  }
+}
+
+export async function uploadForTranscription (server: PeerTubeServer, body: Partial<VideoEdit> = {}) {
+  const { uuid } = await server.videos.upload({
+    attributes: {
+      name: 'video',
+      fixture: join('transcription', 'videos', 'the_last_man_on_earth.mp4'),
+      language: undefined,
+
+      ...body
+    }
+  })
+
+  return uuid
+}
diff --git a/packages/tests/src/transcription/levenshtein-distance.spec.ts b/packages/tests/src/transcription/levenshtein-distance.spec.ts
index d4c502a02..63a2c71f2 100644
--- a/packages/tests/src/transcription/levenshtein-distance.spec.ts
+++ b/packages/tests/src/transcription/levenshtein-distance.spec.ts
@@ -1,5 +1,5 @@
+import { levenshteinDistance } from '@peertube/peertube-transcription-devtools'
 import { expect } from 'chai'
-import { levenshteinDistance } from '@peertube/peertube-transcription'
 
 describe('Levenshtein distance', function () {
   it(`equals 1 when there is only one character difference`, function () {
diff --git a/packages/tests/src/transcription/transcriber-factory.spec.ts b/packages/tests/src/transcription/transcriber-factory.spec.ts
index cd9d9f29f..139b557f8 100644
--- a/packages/tests/src/transcription/transcriber-factory.spec.ts
+++ b/packages/tests/src/transcription/transcriber-factory.spec.ts
@@ -1,17 +1,16 @@
-import { transcriberFactory } from '@peertube/peertube-transcription'
+import { TranscriptionEngineName, transcriberFactory } from '@peertube/peertube-transcription'
+import { createLogger } from 'winston'
 
 describe('Transcriber factory', function () {
-  const transcribers = [
-    'openai-whisper',
-    'whisper-ctranslate2',
-    'whisper-timestamped'
-  ]
+  const transcribers: TranscriptionEngineName[] = [ 'openai-whisper', 'whisper-ctranslate2' ]
 
   describe('Should be able to create a transcriber for each available transcription engine', function () {
-    transcribers.forEach(function (transcriberName) {
+
+    for (const transcriberName of transcribers) {
       it(`Should be able to create a(n) ${transcriberName} transcriber`, function () {
-        transcriberFactory.createFromEngineName(transcriberName)
+        transcriberFactory.createFromEngineName({ engineName: transcriberName, logger: createLogger() })
       })
-    })
+    }
+
   })
 })
diff --git a/packages/tests/src/transcription/transcript/transcript-file-evaluator.spec.ts b/packages/tests/src/transcription/transcript/transcript-file-evaluator.spec.ts
index 93df638db..0ee2fba2f 100644
--- a/packages/tests/src/transcription/transcript/transcript-file-evaluator.spec.ts
+++ b/packages/tests/src/transcription/transcript/transcript-file-evaluator.spec.ts
@@ -1,17 +1,18 @@
 /* eslint-disable @typescript-eslint/no-unused-expressions, no-new, max-len */
-import { TranscriptFile, TranscriptFileEvaluator } from '@peertube/peertube-transcription'
 import { buildAbsoluteFixturePath } from '@peertube/peertube-node-utils'
-import { join } from 'node:path'
-import { mkdir, rm } from 'node:fs/promises'
-import { tmpdir } from 'node:os'
+import { TranscriptFile } from '@peertube/peertube-transcription'
+import { TranscriptFileEvaluator } from '@peertube/peertube-transcription-devtools'
 import { expect } from 'chai'
+import { ensureDir, remove } from 'fs-extra/esm'
+import { tmpdir } from 'node:os'
+import { join } from 'node:path'
 
 describe('Transcript File Evaluator', function () {
   const transcriptDirectory = join(tmpdir(), 'peertube-transcription', 'transcript-file-evaluator')
   const referenceTranscriptFilePath = buildAbsoluteFixturePath('transcription/videos/communiquer-lors-dune-classe-transplantee.txt')
 
   before(async function () {
-    await mkdir(transcriptDirectory, { recursive: true })
+    await ensureDir(transcriptDirectory)
   })
 
   it(`may not compare files in another format than txt`, async function () {
@@ -62,6 +63,6 @@ Ensuite, il pourront lire et commenter ce de leurs camarades ou répondre aux co
   })
 
   after(async function () {
-    await rm(transcriptDirectory, { recursive: true, force: true })
+    await remove(transcriptDirectory)
   })
 })
diff --git a/packages/tests/src/transcription/transcript/transcript-file.spec.ts b/packages/tests/src/transcription/transcript/transcript-file.spec.ts
index 112c246b3..e2e3c194a 100644
--- a/packages/tests/src/transcription/transcript/transcript-file.spec.ts
+++ b/packages/tests/src/transcription/transcript/transcript-file.spec.ts
@@ -1,15 +1,16 @@
 /* eslint-disable @typescript-eslint/no-unused-expressions */
-import { expect } from 'chai'
-import { join } from 'node:path'
-import { mkdir, rm } from 'node:fs/promises'
-import { TranscriptFile } from '@peertube/peertube-transcription'
-import { tmpdir } from 'node:os'
 import { buildAbsoluteFixturePath } from '@peertube/peertube-node-utils'
+import { TranscriptFile } from '@peertube/peertube-transcription'
+import { expect } from 'chai'
+import { ensureDir, remove } from 'fs-extra/esm'
+import { tmpdir } from 'node:os'
+import { join } from 'node:path'
 
 describe('Transcript File', function () {
   const transcriptFileDirectory = join(tmpdir(), 'peertube-transcription', 'transcript-file')
+
   before(async function () {
-    await mkdir(transcriptFileDirectory, { recursive: true })
+    await ensureDir(transcriptFileDirectory)
   })
 
   it(`may creates a new transcript file from scratch`, async function () {
@@ -33,12 +34,11 @@ describe('Transcript File', function () {
   })
 
   it(`fails when loading a file which is obviously not a transcript`, function () {
-
     expect(() => TranscriptFile.fromPath(buildAbsoluteFixturePath('transcription/videos/the_last_man_on_earth.mp4'), 'en'))
       .to.throw(`Couldn't guess transcript format from extension "mp4". Valid formats are: txt, vtt, srt.`)
   })
 
   after(async function () {
-    await rm(transcriptFileDirectory, { recursive: true, force: true })
+    await remove(transcriptFileDirectory)
   })
 })
diff --git a/packages/tests/src/transcription/transcription-run.spec.ts b/packages/tests/src/transcription/transcription-run.spec.ts
deleted file mode 100644
index d877fbded..000000000
--- a/packages/tests/src/transcription/transcription-run.spec.ts
+++ /dev/null
@@ -1 +0,0 @@
-describe('Transcription run', function () {})
diff --git a/packages/tests/src/transcription/utils.spec.ts b/packages/tests/src/transcription/utils.spec.ts
index 487b9aeda..1498169f8 100644
--- a/packages/tests/src/transcription/utils.spec.ts
+++ b/packages/tests/src/transcription/utils.spec.ts
@@ -1,14 +1,16 @@
-import { cp, lstat, mkdir, rm } from 'node:fs/promises'
-import { join } from 'node:path'
-import { tmpdir } from 'node:os'
-import { expect } from 'chai'
-import { downloadFile, unzip } from '@peertube/peertube-transcription'
 import { buildAbsoluteFixturePath } from '@peertube/peertube-node-utils'
+import { downloadFile, unzip } from '@peertube/peertube-transcription-devtools'
+import { expect } from 'chai'
+import { ensureDir, remove } from 'fs-extra/esm'
+import { cp, lstat } from 'node:fs/promises'
+import { tmpdir } from 'node:os'
+import { join } from 'node:path'
 
 describe('downloadFile', function () {
   const testDirectory = join(tmpdir(), 'peertube-transcription', 'utils')
+
   before(async function () {
-    await mkdir(testDirectory, { recursive: true })
+    await ensureDir(testDirectory)
   })
 
   it(`Downloads a file and write it to the disk `, async function () {
@@ -18,7 +20,7 @@ describe('downloadFile', function () {
   })
 
   after(async function () {
-    await rm(testDirectory, { recursive: true, force: true })
+    await remove(testDirectory)
   })
 })
 
@@ -26,8 +28,9 @@ describe('unzip', function () {
   const zipFixtureFileName = 'hello_world.zip'
   const zipFixtureFilePath = buildAbsoluteFixturePath(`transcription/${zipFixtureFileName}`)
   const testDirectory = join(tmpdir(), 'peertube-transcription', 'utils')
+
   before(async function () {
-    await mkdir(testDirectory, { recursive: true })
+    await ensureDir(testDirectory)
   })
 
   it(`Extract zip archive to directory`, async function () {
@@ -39,6 +42,6 @@ describe('unzip', function () {
   })
 
   after(async function () {
-    await rm(testDirectory, { recursive: true, force: true })
+    await remove(testDirectory)
   })
 })
diff --git a/packages/tests/src/transcription/whisper/transcriber/openai-transcriber.spec.ts b/packages/tests/src/transcription/whisper/openai-transcriber.spec.ts
similarity index 70%
rename from packages/tests/src/transcription/whisper/transcriber/openai-transcriber.spec.ts
rename to packages/tests/src/transcription/whisper/openai-transcriber.spec.ts
index d3f3f9ca7..20b2b56d5 100644
--- a/packages/tests/src/transcription/whisper/transcriber/openai-transcriber.spec.ts
+++ b/packages/tests/src/transcription/whisper/openai-transcriber.spec.ts
@@ -1,57 +1,59 @@
 /* eslint-disable @typescript-eslint/no-unused-expressions, max-len */
-import { expect, config } from 'chai'
-import { createLogger } from 'winston'
-import { join } from 'node:path'
-import { mkdir, rm } from 'node:fs/promises'
-import { tmpdir } from 'node:os'
 import { buildAbsoluteFixturePath } from '@peertube/peertube-node-utils'
 import {
-  downloadFile,
-  levenshteinDistance,
   OpenaiTranscriber,
   TranscriptFile,
-  TranscriptFileEvaluator,
   TranscriptionModel,
-  unzip,
   WhisperBuiltinModel
 } from '@peertube/peertube-transcription'
-import { FIXTURE_URLS } from '@tests/shared/fixture-urls.js'
+import { TranscriptFileEvaluator, levenshteinDistance } from '@peertube/peertube-transcription-devtools'
+import { downloadCustomModelsIfNeeded, getCustomModelPath } from '@tests/shared/transcription.js'
+import { config, expect } from 'chai'
+import { ensureDir, remove } from 'fs-extra/esm'
+import { tmpdir } from 'node:os'
+import { join } from 'node:path'
+import { createLogger } from 'winston'
 
 config.truncateThreshold = 0
 
 describe('Open AI Whisper transcriber', function () {
   const tmpDirectory = join(tmpdir(), 'peertube-transcription')
   const transcriptDirectory = join(tmpDirectory, 'transcriber', 'openai')
-  const modelsDirectory = join(tmpDirectory, 'models')
+
   const shortVideoPath = buildAbsoluteFixturePath('transcription/videos/the_last_man_on_earth.mp4')
   const frVideoPath = buildAbsoluteFixturePath('transcription/videos/derive_sectaire.mp4')
+
   const referenceTranscriptFile = new TranscriptFile({
     path: buildAbsoluteFixturePath('transcription/videos/derive_sectaire.txt'),
     language: 'fr',
     format: 'txt'
   })
-  const transcriber = new OpenaiTranscriber(
-    {
+
+  const transcriber = new OpenaiTranscriber({
+    engine: {
       name: 'openai-whisper',
-      requirements: [],
       type: 'binary',
-      binary: 'whisper',
+      command: 'whisper',
       supportedModelFormats: [ 'PyTorch' ],
-      languageDetection: true
+      languageDetection: true,
+      version: ''
     },
-    createLogger(),
-    transcriptDirectory
-  )
+    logger: createLogger()
+  })
+  const model = new TranscriptionModel('tiny')
 
   before(async function () {
-    this.timeout(1 * 1000 * 60)
-    await mkdir(transcriptDirectory, { recursive: true })
-    await unzip(await downloadFile(FIXTURE_URLS.transcriptionModels, tmpDirectory))
+    this.timeout(120000)
+
+    await ensureDir(transcriptDirectory)
+
+    await downloadCustomModelsIfNeeded('tiny.pt')
   })
 
-  it('Should transcribe a media file and provide a valid path to a transcript file in `vtt` format by default', async function () {
+  it('Should transcribe a media file and provide a valid path to a transcript file in `vtt` format', async function () {
     this.timeout(3 * 1000 * 60)
-    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en' })
+
+    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en', format: 'vtt', model, transcriptDirectory })
 
     expect(transcript.format).to.equals('vtt')
     expect(transcript.language).to.equals('en')
@@ -59,7 +61,7 @@ describe('Open AI Whisper transcriber', function () {
   })
 
   it('May produce a transcript file in the `srt` format', async function () {
-    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en', format: 'srt' })
+    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en', format: 'srt', model, transcriptDirectory })
 
     expect(transcript.format).to.equals('srt')
     expect(transcript.language).to.equals('en')
@@ -67,7 +69,7 @@ describe('Open AI Whisper transcriber', function () {
   })
 
   it('May produce a transcript file in the `txt` format', async function () {
-    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en', format: 'txt' })
+    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en', format: 'txt', model, transcriptDirectory })
 
     expect(transcript.format).to.equals('txt')
     expect(transcript.language).to.equals('en')
@@ -80,16 +82,20 @@ describe('Open AI Whisper transcriber', function () {
 
   it('May transcribe a media file using a local PyTorch model', async function () {
     this.timeout(2 * 1000 * 60)
+
     await transcriber.transcribe({
       mediaFilePath: shortVideoPath,
-      model: await TranscriptionModel.fromPath(join(modelsDirectory, 'tiny.pt')),
-      language: 'en'
+      model: await TranscriptionModel.fromPath(getCustomModelPath('tiny.pt')),
+      language: 'en',
+      format: 'vtt',
+      transcriptDirectory
     })
   })
 
   it('May transcribe a media file in french', async function () {
     this.timeout(3 * 1000 * 60)
-    const transcript = await transcriber.transcribe({ mediaFilePath: frVideoPath, language: 'fr', format: 'txt' })
+
+    const transcript = await transcriber.transcribe({ mediaFilePath: frVideoPath, language: 'fr', format: 'txt', model, transcriptDirectory })
 
     expect(transcript.format).to.equals('txt')
     expect(transcript.language).to.equals('fr')
@@ -98,18 +104,21 @@ describe('Open AI Whisper transcriber', function () {
 
   it('Guesses the video language if not provided', async function () {
     this.timeout(3 * 1000 * 60)
-    const transcript = await transcriber.transcribe({ mediaFilePath: frVideoPath })
+
+    const transcript = await transcriber.transcribe({ mediaFilePath: frVideoPath, model, format: 'vtt', transcriptDirectory })
 
     expect(transcript.language).to.equals('fr')
   })
 
-  it('May transcribe a media file in french with small model', async function () {
+  it('May transcribe a media file in french with small model (can be long)', async function () {
     this.timeout(6 * 1000 * 60)
+
     const transcript = await transcriber.transcribe({
       mediaFilePath: frVideoPath,
       language: 'fr',
       format: 'txt',
-      model: new WhisperBuiltinModel('small')
+      model: new WhisperBuiltinModel('small'),
+      transcriptDirectory
     })
 
     expect(transcript.language).to.equals('fr')
@@ -120,6 +129,6 @@ describe('Open AI Whisper transcriber', function () {
   })
 
   after(async function () {
-    await rm(transcriptDirectory, { recursive: true, force: true })
+    await remove(transcriptDirectory)
   })
 })
diff --git a/packages/tests/src/transcription/whisper/transcriber/timestamped-transcriber.spec.ts b/packages/tests/src/transcription/whisper/transcriber/timestamped-transcriber.spec.ts
deleted file mode 100644
index bf4adee9e..000000000
--- a/packages/tests/src/transcription/whisper/transcriber/timestamped-transcriber.spec.ts
+++ /dev/null
@@ -1,133 +0,0 @@
-/* eslint-disable @typescript-eslint/no-unused-expressions, max-len */
-import { expect, config } from 'chai'
-import { createLogger } from 'winston'
-import { join } from 'node:path'
-import { mkdir, rm } from 'node:fs/promises'
-import { tmpdir } from 'node:os'
-import { buildAbsoluteFixturePath } from '@peertube/peertube-node-utils'
-import {
-  OpenaiTranscriber,
-  WhisperTimestampedTranscriber,
-  TranscriptFileEvaluator,
-  TranscriptionModel,
-  WhisperTranscribeArgs,
-  levenshteinDistance, downloadFile, unzip
-} from '@peertube/peertube-transcription'
-import { FIXTURE_URLS } from '@tests/shared/fixture-urls.js'
-
-config.truncateThreshold = 0
-
-describe('Linto timestamped Whisper transcriber', function () {
-  const tmpDirectory = join(tmpdir(), 'peertube-transcription')
-  const transcriptDirectory = join(tmpDirectory, 'transcriber', 'timestamped')
-  const modelsDirectory = join(tmpDirectory, 'models')
-  const shortVideoPath = buildAbsoluteFixturePath('transcription/videos/the_last_man_on_earth.mp4')
-  const frVideoPath = buildAbsoluteFixturePath('transcription/videos/derive_sectaire.mp4')
-  const transcriber = new WhisperTimestampedTranscriber(
-    {
-      name: 'whisper-timestamped',
-      requirements: [],
-      type: 'binary',
-      binary: 'whisper_timestamped',
-      supportedModelFormats: [ 'PyTorch' ],
-      languageDetection: true
-    },
-    createLogger(),
-    transcriptDirectory
-  )
-
-  before(async function () {
-    this.timeout(1 * 1000 * 60)
-    await mkdir(transcriptDirectory, { recursive: true })
-    await unzip(await downloadFile(FIXTURE_URLS.transcriptionModels, tmpDirectory))
-  })
-
-  it('Should transcribe a media file and provide a valid path to a transcript file in `vtt` format by default', async function () {
-    this.timeout(1 * 1000 * 60)
-    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en' })
-
-    expect(transcript.format).to.equals('vtt')
-    expect(transcript.language).to.equals('en')
-    expect(await transcript.read()).not.to.be.empty
-  })
-
-  it('May produce a transcript file in the `srt` format with a ms precision', async function () {
-    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en', format: 'srt' })
-
-    expect(transcript.format).to.equals('srt')
-    expect(transcript.language).to.equals('en')
-    expect(await transcript.read()).not.to.be.empty
-  })
-
-  it('May produce a transcript file in `txt` format', async function () {
-    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en', format: 'txt' })
-
-    expect(transcript.format).to.equals('txt')
-    expect(transcript.language).to.equals('en')
-    expect(await transcript.read()).not.to.be.empty
-    expect(levenshteinDistance(
-      (await transcript.read()).toString(),
-      'December 1965, is that all it has been since I inherited the world only three years, seems like a hundred million.'
-    )).to.be.below(10)
-  })
-
-  it('May transcribe a media file using a local PyTorch model file', async function () {
-    this.timeout(2 * 1000 * 60)
-    await transcriber.transcribe({
-      mediaFilePath: shortVideoPath,
-      model: await TranscriptionModel.fromPath(join(modelsDirectory, 'tiny.pt')),
-      language: 'en'
-    })
-  })
-
-  it('May transcribe a media file in french', async function () {
-    this.timeout(2 * 1000 * 60)
-    const transcript = await transcriber.transcribe({
-      mediaFilePath: frVideoPath,
-      language: 'fr',
-      format: 'txt'
-    })
-
-    expect(transcript.format).to.equals('txt')
-    expect(transcript.language).to.equals('fr')
-    expect(await transcript.read()).not.to.be.empty
-  })
-
-  it('Guesses the video language if not provided', async function () {
-    this.timeout(2 * 1000 * 60)
-    const transcript = await transcriber.transcribe({ mediaFilePath: frVideoPath })
-    expect(transcript.language).to.equals('fr')
-  })
-
-  it('Should produce a text transcript similar to openai-whisper implementation', async function () {
-    this.timeout(11 * 1000 * 60)
-    const transcribeArgs: WhisperTranscribeArgs = {
-      mediaFilePath: frVideoPath,
-      model: await TranscriptionModel.fromPath(join(modelsDirectory, 'tiny.pt')),
-      language: 'fr',
-      format: 'txt'
-    }
-    const transcript = await transcriber.transcribe(transcribeArgs)
-
-    const openaiTranscriber = new OpenaiTranscriber(
-      {
-        name: 'openai-whisper',
-        requirements: [],
-        type: 'binary',
-        binary: 'whisper',
-        supportedModelFormats: [ 'PyTorch' ]
-      },
-      createLogger(),
-      join(transcriptDirectory, 'openai-whisper')
-    )
-    const openaiTranscript = await openaiTranscriber.transcribe(transcribeArgs)
-
-    const transcriptFileEvaluator = new TranscriptFileEvaluator(openaiTranscript, transcript)
-    expect(await transcriptFileEvaluator.wer()).to.be.below(25 / 100)
-    expect(await transcriptFileEvaluator.cer()).to.be.below(15 / 100)
-  })
-
-  after(async function () {
-    await rm(transcriptDirectory, { recursive: true, force: true })
-  })
-})
diff --git a/packages/tests/src/transcription/whisper/transcriber/whisper-ctranslate2.spec.ts b/packages/tests/src/transcription/whisper/whisper-ctranslate2.spec.ts
similarity index 61%
rename from packages/tests/src/transcription/whisper/transcriber/whisper-ctranslate2.spec.ts
rename to packages/tests/src/transcription/whisper/whisper-ctranslate2.spec.ts
index 3a9984e87..cd4f30a39 100644
--- a/packages/tests/src/transcription/whisper/transcriber/whisper-ctranslate2.spec.ts
+++ b/packages/tests/src/transcription/whisper/whisper-ctranslate2.spec.ts
@@ -1,50 +1,58 @@
 /* eslint-disable @typescript-eslint/no-unused-expressions, max-len */
-import { expect, config } from 'chai'
-import { createLogger } from 'winston'
-import { join } from 'node:path'
-import { mkdir, rm } from 'node:fs/promises'
-import { tmpdir } from 'node:os'
 import { buildAbsoluteFixturePath } from '@peertube/peertube-node-utils'
 import {
-  Ctranslate2Transcriber, downloadFile,
-  levenshteinDistance,
+  Ctranslate2Transcriber,
   OpenaiTranscriber,
   TranscriptFile,
-  TranscriptFileEvaluator,
-  TranscriptionModel, unzip,
-  WhisperTranscribeArgs
+  TranscriptionModel
 } from '@peertube/peertube-transcription'
-import { FIXTURE_URLS } from '@tests/shared/fixture-urls.js'
+import { TranscriptFileEvaluator, levenshteinDistance } from '@peertube/peertube-transcription-devtools'
+import { downloadCustomModelsIfNeeded, getCustomModelPath } from '@tests/shared/transcription.js'
+import { config, expect } from 'chai'
+import { ensureDir, remove } from 'fs-extra/esm'
+import { tmpdir } from 'node:os'
+import { join } from 'node:path'
+import { createLogger } from 'winston'
 
 config.truncateThreshold = 0
 
 describe('Whisper CTranslate2 transcriber', function () {
   const tmpDirectory = join(tmpdir(), 'peertube-transcription')
   const transcriptDirectory = join(tmpDirectory, 'transcriber', 'ctranslate2')
-  const modelsDirectory = join(tmpDirectory, 'models')
+
   const shortVideoPath = buildAbsoluteFixturePath('transcription/videos/the_last_man_on_earth.mp4')
   const frVideoPath = buildAbsoluteFixturePath('transcription/videos/derive_sectaire.mp4')
-  const transcriber = new Ctranslate2Transcriber(
-    {
-      name: 'anyNameShouldBeFineReally',
-      requirements: [],
-      type: 'binary',
-      binary: 'whisper-ctranslate2',
-      supportedModelFormats: [],
-      languageDetection: true
-    },
-    createLogger(),
-    transcriptDirectory
-  )
 
-  before(async function () {
-    this.timeout(1 * 1000 * 60)
-    await mkdir(transcriptDirectory, { recursive: true })
-    await unzip(await downloadFile(FIXTURE_URLS.transcriptionModels, tmpDirectory))
+  const transcriber = new Ctranslate2Transcriber({
+    engine: {
+      name: 'whisper-ctranslate2',
+      type: 'binary',
+      command: 'whisper-ctranslate2',
+      supportedModelFormats: [ 'CTranslate2' ],
+      languageDetection: true,
+      version: '0.4.4'
+    },
+    logger: createLogger()
   })
 
-  it('Should transcribe a media file and provide a valid path to a transcript file in `vtt` format by default', async function () {
-    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en' })
+  const model = new TranscriptionModel('tiny')
+
+  before(async function () {
+    this.timeout(120000)
+
+    await ensureDir(transcriptDirectory)
+
+    await downloadCustomModelsIfNeeded('faster-whisper-tiny')
+  })
+
+  it('Should transcribe a media file and provide a valid path to a transcript file in `vtt` format', async function () {
+    const transcript = await transcriber.transcribe({
+      mediaFilePath: shortVideoPath,
+      language: 'en',
+      format: 'vtt',
+      model,
+      transcriptDirectory
+    })
 
     expect(transcript.format).to.equals('vtt')
     expect(transcript.language).to.equals('en')
@@ -52,7 +60,13 @@ describe('Whisper CTranslate2 transcriber', function () {
   })
 
   it('May produce a transcript file in the `srt` format', async function () {
-    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en', format: 'srt' })
+    const transcript = await transcriber.transcribe({
+      mediaFilePath: shortVideoPath,
+      language: 'en',
+      format: 'srt',
+      model,
+      transcriptDirectory
+    })
 
     expect(transcript.format).to.equals('srt')
     expect(transcript.language).to.equals('en')
@@ -60,7 +74,13 @@ describe('Whisper CTranslate2 transcriber', function () {
   })
 
   it('May produce a transcript file in the `txt` format', async function () {
-    const transcript = await transcriber.transcribe({ mediaFilePath: shortVideoPath, language: 'en', format: 'txt' })
+    const transcript = await transcriber.transcribe({
+      mediaFilePath: shortVideoPath,
+      language: 'en',
+      format: 'txt',
+      model,
+      transcriptDirectory
+    })
     expect(await transcript.equals(new TranscriptFile({
       path: join(transcriptDirectory, 'the_last_man_on_earth.txt'),
       format: 'txt',
@@ -78,10 +98,12 @@ describe('Whisper CTranslate2 transcriber', function () {
 
   it('May transcribe a media file using a local CTranslate2 model', async function () {
     this.timeout(2 * 1000 * 60)
+
     const transcript = await transcriber.transcribe({
       mediaFilePath: shortVideoPath,
-      model: await TranscriptionModel.fromPath(join(modelsDirectory, 'faster-whisper-tiny')),
+      model: await TranscriptionModel.fromPath(getCustomModelPath('faster-whisper-tiny')),
       language: 'en',
+      transcriptDirectory,
       format: 'txt'
     })
 
@@ -92,7 +114,14 @@ describe('Whisper CTranslate2 transcriber', function () {
 
   it('May transcribe a media file in french', async function () {
     this.timeout(5 * 1000 * 60)
-    const transcript = await transcriber.transcribe({ mediaFilePath: frVideoPath, language: 'fr', format: 'txt' })
+
+    const transcript = await transcriber.transcribe({
+      mediaFilePath: frVideoPath,
+      language: 'fr',
+      format: 'txt',
+      model,
+      transcriptDirectory
+    })
 
     expect(transcript.format).to.equals('txt')
     expect(transcript.language).to.equals('fr')
@@ -101,30 +130,39 @@ describe('Whisper CTranslate2 transcriber', function () {
 
   it('Guesses the video language if not provided', async function () {
     this.timeout(2 * 1000 * 60)
-    const transcript = await transcriber.transcribe({ mediaFilePath: frVideoPath })
+
+    const transcript = await transcriber.transcribe({ mediaFilePath: frVideoPath, format: 'vtt', model, transcriptDirectory })
     expect(transcript.language).to.equals('fr')
   })
 
   it('Should produce a text transcript similar to openai-whisper implementation', async function () {
     this.timeout(10 * 1000 * 60)
-    const transcribeArgs: WhisperTranscribeArgs = {
+
+    const transcribeArgs = {
       mediaFilePath: frVideoPath,
       language: 'fr',
-      format: 'txt'
+      format: 'txt' as 'txt',
+      transcriptDirectory,
+      model
     }
+
     const transcript = await transcriber.transcribe(transcribeArgs)
-    const openaiTranscriber = new OpenaiTranscriber(
-      {
+
+    const openaiTranscriber = new OpenaiTranscriber({
+      engine: {
         name: 'openai-whisper',
-        requirements: [],
         type: 'binary',
-        binary: 'whisper',
-        supportedModelFormats: [ 'PyTorch' ]
+        command: 'whisper',
+        supportedModelFormats: [ 'PyTorch' ],
+        version: '0.4.4'
       },
-      createLogger(),
-      join(transcriptDirectory, 'openai-whisper')
-    )
-    const openaiTranscript = await openaiTranscriber.transcribe(transcribeArgs)
+      logger: createLogger()
+    })
+    const openaiTranscript = await openaiTranscriber.transcribe({
+      ...transcribeArgs,
+
+      transcriptDirectory: join(transcriptDirectory, 'openai-whisper')
+    })
 
     const transcriptFileEvaluator = new TranscriptFileEvaluator(openaiTranscript, transcript)
     expect(await transcriptFileEvaluator.wer()).to.be.below(20 / 100)
@@ -132,6 +170,6 @@ describe('Whisper CTranslate2 transcriber', function () {
   })
 
   after(async function () {
-    await rm(transcriptDirectory, { recursive: true, force: true })
+    await remove(transcriptDirectory)
   })
 })
diff --git a/packages/tests/tsconfig.json b/packages/tests/tsconfig.json
index 5b8e2a6c4..642b97393 100644
--- a/packages/tests/tsconfig.json
+++ b/packages/tests/tsconfig.json
@@ -8,13 +8,13 @@
       "@tests/*": [ "./src/*" ],
       "@server/*": [ "../../server/core/*" ],
       "@peertube/peertube-transcription": [ "../transcription" ],
-      "@peertube/peertube-jiwer": [ "../jiwer" ],
+      "@peertube/peertube-transcription-devtools": [ "../transcription-devtools" ],
     }
   },
   "references": [
     { "path": "../core-utils" },
     { "path": "../ffmpeg" },
-    { "path": "../jiwer" },
+    { "path": "../transcription-devtools" },
     { "path": "../models" },
     { "path": "../node-utils" },
     { "path": "../typescript-utils" },
diff --git a/packages/transcription-devtools/README.md b/packages/transcription-devtools/README.md
new file mode 100644
index 000000000..9e9799ccf
--- /dev/null
+++ b/packages/transcription-devtools/README.md
@@ -0,0 +1,63 @@
+# Transcription DevTools
+
+Includes:
+  * __JiWER__ CLI NodeJS wrapper
+  * Benchmark tool to test multiple transcription engines
+  * TypeScript classes to evaluate word-error-rate of files generated by the transcription
+
+## Build
+
+```sh
+npm run build
+```
+
+## Benchmark
+
+A benchmark of available __transcribers__ might be run with:
+```sh
+npm run benchmark
+```
+```
+┌────────────────────────┬───────────────────────┬───────────────────────┬──────────┬────────┬───────────────────────┐
+│        (index)         │          WER          │          CER          │ duration │ model  │        engine         │
+├────────────────────────┼───────────────────────┼───────────────────────┼──────────┼────────┼───────────────────────┤
+│ 5yZGBYqojXe7nuhq1TuHvz │ '28.39506172839506%'  │  '9.62457337883959%'  │  '41s'   │ 'tiny' │   'openai-whisper'    │
+│ x6qREJ2AkTU4e5YmvfivQN │ '29.75206611570248%'  │ '10.46195652173913%'  │  '15s'   │ 'tiny' │ 'whisper-ctranslate2' │
+└────────────────────────┴───────────────────────┴───────────────────────┴──────────┴────────┴───────────────────────┘
+```
+
+The benchmark may be run with multiple model builtin sizes:
+
+```sh
+MODELS=tiny,small,large npm run benchmark
+```
+
+## Jiwer
+
+> *JiWER is a python tool for computing the word-error-rate of ASR systems.*
+> https://jitsi.github.io/jiwer/cli/
+
+__JiWER__ serves as a reference implementation to calculate errors rates between 2 text files:
+- WER (Word Error Rate)
+- CER (Character Error Rate)
+
+
+### Usage
+
+```typescript
+const jiwerCLI = new JiwerClI('./reference.txt', './hypothesis.txt')
+
+// WER as a percentage, ex: 0.03 -> 3%
+console.log(await jiwerCLI.wer())
+
+// CER as a percentage: 0.01 -> 1%
+console.log(await jiwerCLI.cer())
+
+// Detailed comparison report
+console.log(await jiwerCLI.alignment())
+```
+
+## Resources
+
+- https://jitsi.github.io/jiwer/
+- https://github.com/rapidfuzz/RapidFuzz
diff --git a/packages/jiwer/package.json b/packages/transcription-devtools/package.json
similarity index 65%
rename from packages/jiwer/package.json
rename to packages/transcription-devtools/package.json
index b01476956..8856fb782 100644
--- a/packages/jiwer/package.json
+++ b/packages/transcription-devtools/package.json
@@ -1,5 +1,5 @@
 {
-  "name": "@peertube/peertube-jiwer",
+  "name": "@peertube/peertube-transcription-devtools",
   "private": true,
   "version": "0.0.0",
   "main": "dist/index.js",
@@ -12,9 +12,9 @@
   "type": "module",
   "devDependencies": {},
   "scripts": {
-    "preinstall": "pip install -r requirements.txt",
     "build": "tsc",
-    "watch": "tsc -w"
+    "watch": "tsc -w",
+    "benchmark": "tsx --conditions=peertube:tsx --tsconfig ./tsconfig.json ./src/benchmark.ts"
   },
   "dependencies": {}
 }
diff --git a/packages/transcription-devtools/requirements.txt b/packages/transcription-devtools/requirements.txt
new file mode 100644
index 000000000..68a1e7c5b
--- /dev/null
+++ b/packages/transcription-devtools/requirements.txt
@@ -0,0 +1 @@
+jiwer
diff --git a/packages/transcription/src/benchmark.ts b/packages/transcription-devtools/src/benchmark.ts
similarity index 82%
rename from packages/transcription/src/benchmark.ts
rename to packages/transcription-devtools/src/benchmark.ts
index 4cac16449..884af03c0 100644
--- a/packages/transcription/src/benchmark.ts
+++ b/packages/transcription-devtools/src/benchmark.ts
@@ -1,17 +1,18 @@
-import { createLogger, transports, format } from 'winston'
-import { join } from 'node:path'
-import { performance, PerformanceObserver } from 'node:perf_hooks'
-import { tmpdir } from 'node:os'
-import { rm, mkdir } from 'node:fs/promises'
-import { buildAbsoluteFixturePath, buildSUUID, SUUID } from '@peertube/peertube-node-utils'
-import {
-  transcriberFactory,
-  TranscriptFile,
-  TranscriptFileEvaluator,
-  TranscriptionEngine,
-  TranscriptionModel
-} from '@peertube/peertube-transcription'
 import { millisecondsToTime } from '@peertube/peertube-core-utils'
+import { SUUID, buildAbsoluteFixturePath, buildSUUID } from '@peertube/peertube-node-utils'
+import {
+  TranscriptFile,
+  TranscriptionEngine,
+  TranscriptionEngineName,
+  TranscriptionModel,
+  transcriberFactory
+} from '@peertube/peertube-transcription'
+import { ensureDir, remove } from 'fs-extra/esm'
+import { tmpdir } from 'node:os'
+import { join } from 'node:path'
+import { PerformanceObserver, performance } from 'node:perf_hooks'
+import { createLogger, format, transports } from 'winston'
+import { TranscriptFileEvaluator } from './transcript-file-evaluator.js'
 
 interface BenchmarkResult {
   uuid: SUUID
@@ -59,16 +60,15 @@ const formatBenchmarkResult = ({ WER, CER, duration, engine, model }: Partial<Be
 void (async () => {
   const logger = createLogger()
   logger.add(new transports.Console({ format: format.printf(log => log.message) }))
-  const transcribers = [
-    'openai-whisper',
-    'whisper-ctranslate2',
-    'whisper-timestamped'
-  ]
+
+  const transcribers: TranscriptionEngineName[] = [ 'openai-whisper', 'whisper-ctranslate2' ]
   const models = process.env.MODELS
     ? process.env.MODELS.trim().split(',').map(modelName => modelName.trim()).filter(modelName => modelName)
     : [ 'tiny' ]
 
   const transcriptDirectory = join(tmpdir(), 'peertube-transcription', 'benchmark')
+  const pipDirectory = join(tmpdir(), 'peertube-transcription', 'pip')
+
   const mediaFilePath = buildAbsoluteFixturePath('transcription/videos/derive_sectaire.mp4')
   const referenceTranscriptFile = new TranscriptFile({
     path: buildAbsoluteFixturePath('transcription/videos/derive_sectaire.txt'),
@@ -79,7 +79,7 @@ void (async () => {
   let benchmarkResults: Record<string, BenchmarkResult> = {}
 
   // before
-  await mkdir(transcriptDirectory, { recursive: true })
+  await ensureDir(transcriptDirectory)
   const performanceObserver = new PerformanceObserver((items) => {
     items
       .getEntries()
@@ -97,11 +97,13 @@ void (async () => {
   for (const transcriberName of transcribers) {
     logger.info(`Create "${transcriberName}" transcriber for the benchmark...`)
 
-    const transcriber = transcriberFactory.createFromEngineName(
-      transcriberName,
-      createLogger(),
-      transcriptDirectory
-    )
+    const transcriber = transcriberFactory.createFromEngineName({
+      engineName: transcriberName,
+      logger: createLogger(),
+      binDirectory: join(pipDirectory, 'bin')
+    })
+
+    await transcriber.install(pipDirectory)
 
     for (const modelName of models) {
       logger.info(`Run benchmark with "${modelName}" model:`)
@@ -110,6 +112,7 @@ void (async () => {
       const transcriptFile = await transcriber.transcribe({
         mediaFilePath,
         model,
+        transcriptDirectory,
         language: 'fr',
         format: 'txt',
         runId: uuid
@@ -134,6 +137,6 @@ void (async () => {
   Object.values(benchmarkResultsGroupedByModel).forEach(benchmark => console.table(benchmark))
 
   // after
-  await rm(transcriptDirectory, { recursive: true, force: true })
+  await remove(transcriptDirectory)
   performance.clearMarks()
 })()
diff --git a/packages/transcription-devtools/src/index.ts b/packages/transcription-devtools/src/index.ts
new file mode 100644
index 000000000..fd7f80b9b
--- /dev/null
+++ b/packages/transcription-devtools/src/index.ts
@@ -0,0 +1,5 @@
+export * from './jiwer-cli.js'
+export * from './levenshtein.js'
+export * from './transcript-file-evaluator-interface.js'
+export * from './transcript-file-evaluator.js'
+export * from './utils.js'
diff --git a/packages/jiwer/src/jiwer-cli.ts b/packages/transcription-devtools/src/jiwer-cli.ts
similarity index 100%
rename from packages/jiwer/src/jiwer-cli.ts
rename to packages/transcription-devtools/src/jiwer-cli.ts
diff --git a/packages/transcription/src/levenshtein.ts b/packages/transcription-devtools/src/levenshtein.ts
similarity index 100%
rename from packages/transcription/src/levenshtein.ts
rename to packages/transcription-devtools/src/levenshtein.ts
diff --git a/packages/transcription/src/transcript/transcript-file-evaluator-interface.ts b/packages/transcription-devtools/src/transcript-file-evaluator-interface.ts
similarity index 100%
rename from packages/transcription/src/transcript/transcript-file-evaluator-interface.ts
rename to packages/transcription-devtools/src/transcript-file-evaluator-interface.ts
diff --git a/packages/transcription/src/transcript/transcript-file-evaluator.ts b/packages/transcription-devtools/src/transcript-file-evaluator.ts
similarity index 75%
rename from packages/transcription/src/transcript/transcript-file-evaluator.ts
rename to packages/transcription-devtools/src/transcript-file-evaluator.ts
index 7ae411ee4..faf696e75 100644
--- a/packages/transcription/src/transcript/transcript-file-evaluator.ts
+++ b/packages/transcription-devtools/src/transcript-file-evaluator.ts
@@ -1,14 +1,14 @@
 import assert from 'node:assert'
-import { JiwerClI } from '@peertube/peertube-jiwer'
 import { TranscriptFileEvaluatorInterface } from './transcript-file-evaluator-interface.js'
-import { TranscriptFileInterface } from './index.js'
+import { TranscriptFile } from '@peertube/peertube-transcription'
+import { JiwerClI } from './jiwer-cli.js'
 
 export class TranscriptFileEvaluator implements TranscriptFileEvaluatorInterface {
-  referenceTranscriptFile: TranscriptFileInterface
-  hypothesisTranscriptFile: TranscriptFileInterface
+  referenceTranscriptFile: TranscriptFile
+  hypothesisTranscriptFile: TranscriptFile
   jiwerCLI: JiwerClI
 
-  constructor (referenceTranscriptFile: TranscriptFileInterface, hypothesisTranscriptFile: TranscriptFileInterface) {
+  constructor (referenceTranscriptFile: TranscriptFile, hypothesisTranscriptFile: TranscriptFile) {
     assert(referenceTranscriptFile.format === 'txt', 'Can only evaluate txt transcript file')
     assert(hypothesisTranscriptFile.format === 'txt', 'Can only evaluate txt transcript file')
 
diff --git a/packages/transcription/src/utils.ts b/packages/transcription-devtools/src/utils.ts
similarity index 100%
rename from packages/transcription/src/utils.ts
rename to packages/transcription-devtools/src/utils.ts
diff --git a/packages/jiwer/tsconfig.json b/packages/transcription-devtools/tsconfig.json
similarity index 51%
rename from packages/jiwer/tsconfig.json
rename to packages/transcription-devtools/tsconfig.json
index 58fa2330b..39f9bd4b6 100644
--- a/packages/jiwer/tsconfig.json
+++ b/packages/transcription-devtools/tsconfig.json
@@ -3,6 +3,9 @@
   "compilerOptions": {
     "outDir": "./dist",
     "rootDir": "src",
-    "tsBuildInfoFile": "./dist/.tsbuildinfo"
-  }
+    "tsBuildInfoFile": "./dist/.tsbuildinfo",
+  },
+  "references": [
+    { "path": "../transcription" }
+  ]
 }
diff --git a/packages/transcription/README.md b/packages/transcription/README.md
index b16ff75d4..7036f54b6 100644
--- a/packages/transcription/README.md
+++ b/packages/transcription/README.md
@@ -1,45 +1,50 @@
 # Transcription
 
 Video **transcription** consists in transcribing the audio content of a video to a text.
+
 > This process might be called __Automatic Speech Recognition__ or __Speech to Text__ in more general context.
 
-Provide a common API to many transcription backend, currently :
-- `openai-whisper` CLI
-- `faster-whisper` (*via* `whisper-ctranslate2` CLI)
-- `whisper-timestamped`
+Provide a common API to many transcription backend, currently:
+  - `openai-whisper` CLI
+  - `faster-whisper` (*via* `whisper-ctranslate2` CLI)
 
 > Potential candidates could be: whisper-cpp, vosk, ...
 
 ## Requirements
-- Python
-- PIP
+  - Python 3
+  - PIP
 
 And at least one of the following transcription backend:
-- Python :
-  - `openai-whisper`
-  - `whisper-ctranslate2>=0.4.3`
-  - `whisper-timestamped>=1.15.4`
-
-And to run the transcript evaluation tests :
-- Python
-  - `jiwer>=3.04`
+  - Python:
+    - `openai-whisper`
+    - `whisper-ctranslate2>=0.4.3`
 
 ## Usage
 
-Create a transcriber manually :
+Create a transcriber manually:
+
 ```typescript
 import { OpenaiTranscriber } from '@peertube/peertube-transcription'
 
 (async () => {
-  // create a transcriber powered by OpeanAI Whisper CLI
+  // Optional if you want to use a local installation of transcribe engines
+  const binDirectory = 'local/pip/path/bin'
+
+  // Create a transcriber powered by OpenAI Whisper CLI
   const transcriber = new OpenaiTranscriber({
     name: 'openai-whisper',
-    binary: 'whisper',
-    languageDetection: true
+    command: 'whisper',
+    languageDetection: true,
+    binDirectory
   });
 
+  // If not installed globally, install the transcriber engine (use pip under the hood)
+  await transcriber.install('local/pip/path')
+
+  // Transcribe
   const transcriptFile = await transcriber.transcribe({
     mediaFilePath: './myVideo.mp4',
+    model: 'tiny',
     format: 'txt'
   });
 
@@ -55,38 +60,24 @@ import { WhisperBuiltinModel } from '@peertube/peertube-transcription/dist'
 
 const transcriptFile = await transcriber.transcribe({
   mediaFilePath: './myVideo.mp4',
-  model: WhisperBuiltinModel.fromPath('./models/large.pt'),
+  model: await WhisperBuiltinModel.fromPath('./models/large.pt'),
   format: 'txt'
 });
 ```
 
 You may use the builtin Factory if you're happy with the default configuration:
+
 ```Typescript
 import { transcriberFactory } from '@peertube/peertube-transcription'
-transcriberFactory.createFromEngineName('openai-whisper')
+
+transcriberFactory.createFromEngineName({
+  engineName: transcriberName,
+  logger: compatibleWinstonLogger,
+  transcriptDirectory: '/tmp/transcription'
+})
 ```
 > For further usage [../tests/src/transcription/whisper/transcriber/openai-transcriber.spec.ts](../tests/src/transcription/whisper/transcriber/openai-transcriber.spec.ts)
 
-## Benchmark
-
-A benchmark of available __transcribers__ might be run with:
-```sh
-npm run benchmark
-```
-```
-┌────────────────────────┬───────────────────────┬───────────────────────┬──────────┬────────┬───────────────────────┐
-│        (index)         │          WER          │          CER          │ duration │ model  │        engine         │
-├────────────────────────┼───────────────────────┼───────────────────────┼──────────┼────────┼───────────────────────┤
-│ 5yZGBYqojXe7nuhq1TuHvz │ '28.39506172839506%'  │  '9.62457337883959%'  │  '41s'   │ 'tiny' │   'openai-whisper'    │
-│ x6qREJ2AkTU4e5YmvfivQN │ '29.75206611570248%'  │ '10.46195652173913%'  │  '15s'   │ 'tiny' │ 'whisper-ctranslate2' │
-│ qbt6BekKMVzxq4KCSLCzt3 │ '31.020408163265305%' │ '10.784982935153584%' │  '20s'   │ 'tiny' │ 'whisper-timestamped' │
-└────────────────────────┴───────────────────────┴───────────────────────┴──────────┴────────┴───────────────────────┘
-```
-
-The benchmark may be run with multiple model builtin sizes:
-```sh
-MODELS=tiny,small,large npm run benchmark
-```
 
 ## Lexicon
 - ONNX: Open Neural Network eXchange. A specification, the ONNX Runtime run these models.
diff --git a/packages/transcription/package.json b/packages/transcription/package.json
index 9138de4dc..366a08686 100644
--- a/packages/transcription/package.json
+++ b/packages/transcription/package.json
@@ -12,10 +12,8 @@
   "type": "module",
   "devDependencies": {},
   "scripts": {
-    "preinstall": "pip install -r requirements.txt",
     "build": "tsc",
-    "watch": "tsc -w",
-    "benchmark": "tsx --conditions=peertube:tsx --tsconfig ./tsconfig.json ./src/benchmark.ts"
+    "watch": "tsc -w"
   },
   "dependencies": {}
 }
diff --git a/packages/transcription/requirements.txt b/packages/transcription/requirements.txt
deleted file mode 100644
index bc457025c..000000000
--- a/packages/transcription/requirements.txt
+++ /dev/null
@@ -1,3 +0,0 @@
-openai-whisper==20231117
-whisper-ctranslate2==0.4.4
-whisper-timestamped==1.15.4
diff --git a/packages/transcription/src/abstract-transcriber.ts b/packages/transcription/src/abstract-transcriber.ts
index b9a0f66e5..b3d998b72 100644
--- a/packages/transcription/src/abstract-transcriber.ts
+++ b/packages/transcription/src/abstract-transcriber.ts
@@ -1,38 +1,47 @@
-import { createLogger, Logger } from 'winston'
-import { join } from 'node:path'
+import { SimpleLogger } from '@peertube/peertube-models'
+import { buildSUUID, SUUID } from '@peertube/peertube-node-utils'
 import { PerformanceObserver } from 'node:perf_hooks'
-import { buildSUUID, SUUID, root } from '@peertube/peertube-node-utils'
+import { join } from 'path'
+import { TranscriptFile, TranscriptFormat } from './transcript-file.js'
 import { TranscriptionEngine } from './transcription-engine.js'
 import { TranscriptionModel } from './transcription-model.js'
 import { TranscriptionRun } from './transcription-run.js'
-import { TranscriptFile, TranscriptFormat } from './transcript/index.js'
 
 export interface TranscribeArgs {
   mediaFilePath: string
   model: TranscriptionModel
+  format: TranscriptFormat
+  transcriptDirectory: string
+
   language?: string
-  format?: TranscriptFormat
   runId?: SUUID
 }
 
 export abstract class AbstractTranscriber {
-  public static DEFAULT_TRANSCRIPT_DIRECTORY = join(root(), 'dist', 'transcripts')
-
   engine: TranscriptionEngine
-  logger: Logger
-  transcriptDirectory: string
-  performanceObserver?: PerformanceObserver
-  run?: TranscriptionRun
 
-  constructor (
-    engine: TranscriptionEngine,
-    logger: Logger = createLogger(),
-    transcriptDirectory: string = AbstractTranscriber.DEFAULT_TRANSCRIPT_DIRECTORY,
+  protected binDirectory: string
+  protected enginePath: string
+
+  protected logger: SimpleLogger
+
+  protected performanceObserver?: PerformanceObserver
+  protected run?: TranscriptionRun
+
+  constructor (options: {
+    engine: TranscriptionEngine
+    binDirectory?: string
+    enginePath?: string
+
+    logger: SimpleLogger
     performanceObserver?: PerformanceObserver
-  ) {
+  }) {
+    const { engine, logger, enginePath, binDirectory, performanceObserver } = options
+
     this.engine = engine
+    this.enginePath = enginePath
     this.logger = logger
-    this.transcriptDirectory = transcriptDirectory
+    this.binDirectory = binDirectory
     this.performanceObserver = performanceObserver
   }
 
@@ -59,11 +68,14 @@ export abstract class AbstractTranscriber {
     return model.format === 'PyTorch'
   }
 
-  abstract transcribe ({
-    mediaFilePath,
-    model,
-    language,
-    format = 'vtt',
-    runId = buildSUUID()
-  }: TranscribeArgs): Promise<TranscriptFile>
+  protected getEngineBinary () {
+    if (this.enginePath) return this.enginePath
+    if (this.binDirectory) return join(this.binDirectory, this.engine.command)
+
+    return this.engine.command
+  }
+
+  abstract transcribe (options: TranscribeArgs): Promise<TranscriptFile>
+
+  abstract install (path: string): Promise<void>
 }
diff --git a/packages/transcription/src/index.ts b/packages/transcription/src/index.ts
index 44eaa0799..9ae289181 100644
--- a/packages/transcription/src/index.ts
+++ b/packages/transcription/src/index.ts
@@ -1,13 +1,12 @@
 import { TranscriberFactory } from './transcriber-factory.js'
 import { engines } from './whisper/index.js'
 
-export * from './transcript/index.js'
-export * from './levenshtein.js'
+export * from './abstract-transcriber.js'
+export * from './transcript-file.js'
 export * from './subtitle.js'
 export * from './transcription-engine.js'
 export * from './transcription-model.js'
 export * from './transcription-run.js'
-export * from './utils.js'
 export * from './whisper/index.js'
 
 export const transcriberFactory = new TranscriberFactory(engines)
diff --git a/packages/transcription/src/transcriber-factory.ts b/packages/transcription/src/transcriber-factory.ts
index cb22d617c..c682052bc 100644
--- a/packages/transcription/src/transcriber-factory.ts
+++ b/packages/transcription/src/transcriber-factory.ts
@@ -1,10 +1,6 @@
-import { Logger, createLogger } from 'winston'
-import { TranscriptionEngine } from './transcription-engine.js'
-import {
-  Ctranslate2Transcriber,
-  OpenaiTranscriber, WhisperTimestampedTranscriber
-} from './whisper/index.js'
-import { AbstractTranscriber } from './abstract-transcriber.js'
+import { SimpleLogger } from '@peertube/peertube-models'
+import { TranscriptionEngine, TranscriptionEngineName } from './transcription-engine.js'
+import { Ctranslate2Transcriber, OpenaiTranscriber } from './whisper/index.js'
 
 export class TranscriberFactory {
   engines: TranscriptionEngine[]
@@ -13,26 +9,28 @@ export class TranscriberFactory {
     this.engines = engines
   }
 
-  createFromEngineName (
-    engineName: string,
-    logger: Logger = createLogger(),
-    transcriptDirectory: string = AbstractTranscriber.DEFAULT_TRANSCRIPT_DIRECTORY
-  ) {
-    const engine = this.getEngineByName(engineName)
+  createFromEngineName (options: {
+    engineName: TranscriptionEngineName
+    enginePath?: string
+    binDirectory?: string
 
-    const transcriberArgs: ConstructorParameters<typeof AbstractTranscriber> = [
-      engine,
-      logger,
-      transcriptDirectory
-    ]
+    logger: SimpleLogger
+  }) {
+    const { engineName } = options
+
+    const transcriberArgs = {
+      ...options,
+
+      engine: this.getEngineByName(engineName)
+    }
 
     switch (engineName) {
       case 'openai-whisper':
-        return new OpenaiTranscriber(...transcriberArgs)
+        return new OpenaiTranscriber(transcriberArgs)
+
       case 'whisper-ctranslate2':
-        return new Ctranslate2Transcriber(...transcriberArgs)
-      case 'whisper-timestamped':
-        return new WhisperTimestampedTranscriber(...transcriberArgs)
+        return new Ctranslate2Transcriber(transcriberArgs)
+
       default:
         throw new Error(`Unimplemented engine ${engineName}`)
     }
diff --git a/packages/transcription/src/transcript/transcript-file.ts b/packages/transcription/src/transcript-file.ts
similarity index 69%
rename from packages/transcription/src/transcript/transcript-file.ts
rename to packages/transcription/src/transcript-file.ts
index 36af9fa8f..b2c17e0b6 100644
--- a/packages/transcription/src/transcript/transcript-file.ts
+++ b/packages/transcription/src/transcript-file.ts
@@ -1,20 +1,16 @@
-import { statSync } from 'node:fs'
+import assert from 'node:assert'
 import { readFile, writeFile } from 'node:fs/promises'
 import { extname } from 'node:path'
-import assert from 'node:assert'
-import { TranscriptFileInterface, TranscriptFormat } from './transcript-file-interface.js'
-import { TranscriptFileEvaluator } from './transcript-file-evaluator.js'
-import { srtToTxt } from '../subtitle.js'
-import { levenshteinDistance } from '../levenshtein.js'
+import { srtToTxt } from './subtitle.js'
 
-export class TranscriptFile implements TranscriptFileInterface {
+export type TranscriptFormat = 'txt' | 'vtt' | 'srt' | 'json'
+
+export class TranscriptFile {
   path: string
   language: string
   format: TranscriptFormat = 'vtt'
 
   constructor ({ path, language, format = 'vtt' }: { path: string, language: string, format?: TranscriptFormat }) {
-    statSync(path)
-
     this.path = path
     this.language = language
     this.format = format
@@ -25,7 +21,7 @@ export class TranscriptFile implements TranscriptFileInterface {
    * @see https://nodejs.org/docs/latest-v18.x/api/fs.html#filehandlereadfileoptions for options
    */
   async read (options: Parameters<typeof readFile>[1] = 'utf8') {
-    return await readFile(this.path, options)
+    return readFile(this.path, options)
   }
 
   static fromPath (path: string, language = 'en') {
@@ -68,21 +64,7 @@ export class TranscriptFile implements TranscriptFileInterface {
     return content === transcriptContent
   }
 
-  cer (transcript: TranscriptFile) {
-    return (new TranscriptFileEvaluator(this, transcript)).cer()
-  }
-
-  async evaluate (transcript: TranscriptFile) {
-    const evaluator = new TranscriptFileEvaluator(this, transcript)
-
-    return evaluator.evaluate()
-  }
-
   async readAsTxt () {
     return srtToTxt(String(await this.read()))
   }
-
-  async distance (transcript: TranscriptFile) {
-    return levenshteinDistance(await this.readAsTxt(), await transcript.readAsTxt())
-  }
 }
diff --git a/packages/transcription/src/transcript/index.ts b/packages/transcription/src/transcript/index.ts
deleted file mode 100644
index bd76d1a86..000000000
--- a/packages/transcription/src/transcript/index.ts
+++ /dev/null
@@ -1,3 +0,0 @@
-export * from './transcript-file.js'
-export * from './transcript-file-evaluator.js'
-export * from './transcript-file-interface.js'
diff --git a/packages/transcription/src/transcript/transcript-file-interface.ts b/packages/transcription/src/transcript/transcript-file-interface.ts
deleted file mode 100644
index d30b6913f..000000000
--- a/packages/transcription/src/transcript/transcript-file-interface.ts
+++ /dev/null
@@ -1,3 +0,0 @@
-export type TranscriptFormat = 'txt' | 'vtt' | 'srt' | 'json'
-
-export type TranscriptFileInterface = { path: string, language?: string, format: TranscriptFormat }
diff --git a/packages/transcription/src/transcription-engine.ts b/packages/transcription/src/transcription-engine.ts
index 3174e3980..ebe81ddfe 100644
--- a/packages/transcription/src/transcription-engine.ts
+++ b/packages/transcription/src/transcription-engine.ts
@@ -1,23 +1,16 @@
 import { ModelFormat } from './transcription-model.js'
 
-/**
- * The engine, or framework.
- */
-export class TranscriptionEngine {
-  name: string
+export type TranscriptionEngineName = 'openai-whisper' | 'whisper-ctranslate2'
+
+export interface TranscriptionEngine {
+  name: TranscriptionEngineName
   description?: string
   language?: string
-  requirements: string[]
-  type: 'binary' | 'bindings' | 'ws'
-  binary: string
+  type: 'binary'
+  command: string
+  version: string
   license?: string
   forgeURL?: string
   supportedModelFormats: ModelFormat[]
   languageDetection?: true
-  // There could be a default models.
-  // There could be a list of default models
-
-  constructor (parameters: TranscriptionEngine) {
-    Object.assign(this, parameters)
-  }
 }
diff --git a/packages/transcription/src/transcription-run.ts b/packages/transcription/src/transcription-run.ts
index 608afbb6a..570027a93 100644
--- a/packages/transcription/src/transcription-run.ts
+++ b/packages/transcription/src/transcription-run.ts
@@ -1,11 +1,12 @@
+import { SimpleLogger } from '@peertube/peertube-models'
 import { buildSUUID, SUUID } from '@peertube/peertube-node-utils'
-import { createLogger, Logger } from 'winston'
+import { createLogger } from 'winston'
 
 export class TranscriptionRun {
   uuid: SUUID
-  logger: Logger
+  logger: SimpleLogger
 
-  constructor (logger = createLogger(), uuid: SUUID = buildSUUID()) {
+  constructor (logger: SimpleLogger = createLogger(), uuid: SUUID = buildSUUID()) {
     this.uuid = uuid
     this.logger = logger
   }
@@ -26,8 +27,8 @@ export class TranscriptionRun {
         this.getStartPerformanceMarkName(),
         this.getEndPerformanceMarkName()
       )
-    } catch (e) {
-      this.logger.log({ level: 'error', message: e })
+    } catch (err) {
+      this.logger.error(err.message, { err })
     }
   }
 
diff --git a/packages/transcription/src/whisper/README.md b/packages/transcription/src/whisper/README.md
deleted file mode 100644
index e69de29bb..000000000
diff --git a/packages/transcription/src/whisper/engines.ts b/packages/transcription/src/whisper/engines.ts
index 711b098b7..ac4ce0bab 100644
--- a/packages/transcription/src/whisper/engines.ts
+++ b/packages/transcription/src/whisper/engines.ts
@@ -1,51 +1,28 @@
 import { TranscriptionEngine } from '../transcription-engine.js'
 
 export const engines: TranscriptionEngine[] = [
-  {
-    name : 'whisper-cpp',
-    description : 'High-performance inference of OpenAI\'s Whisper automatic speech recognition model',
-    type: 'binary',
-    binary: 'main',
-    language : 'cpp',
-    requirements : [],
-    forgeURL : 'https://github.com/ggerganov/whisper.cpp',
-    license : 'MIT',
-    supportedModelFormats: [ 'ONNX' ]
-  },
   {
     name: 'openai-whisper',
     description: 'High-performance inference of OpenAI\'s Whisper automatic speech recognition model',
-    requirements: [ 'python', 'pyTorch', 'ffmpeg' ],
     language: 'python',
     type: 'binary',
-    binary: 'whisper',
+    command: 'whisper',
     forgeURL: 'https://github.com/openai/whisper',
     license: 'MIT',
     supportedModelFormats: [ 'PyTorch' ],
-    languageDetection: true
+    languageDetection: true,
+    version: '20231117'
   },
   {
     name: 'whisper-ctranslate2',
-    description: '',
-    requirements: [ 'python' ],
+    description: 'Whisper command line client compatible with original OpenAI client based on CTranslate2.',
     language: 'python',
     type: 'binary',
-    binary: 'whisper-ctranslate2',
-    forgeURL: 'https://github.com/openai/whisper',
+    command: 'whisper-ctranslate2',
+    forgeURL: 'https://github.com/Softcatala/whisper-ctranslate2',
     license: 'MIT',
     supportedModelFormats: [ 'CTranslate2' ],
-    languageDetection: true
-  },
-  {
-    name: 'whisper-timestamped',
-    description: '',
-    requirements: [ 'python' ],
-    language: 'python',
-    type: 'binary',
-    binary: 'whisper_timestamped',
-    forgeURL: 'https://github.com/openai/whisper',
-    license: 'MIT',
-    supportedModelFormats: [ 'CTranslate2' ],
-    languageDetection: true
+    languageDetection: true,
+    version: '0.4.4'
   }
 ]
diff --git a/packages/transcription/src/whisper/transcriber/ctranslate2-transcriber.ts b/packages/transcription/src/whisper/transcriber/ctranslate2-transcriber.ts
index 01b9739a3..260488123 100644
--- a/packages/transcription/src/whisper/transcriber/ctranslate2-transcriber.ts
+++ b/packages/transcription/src/whisper/transcriber/ctranslate2-transcriber.ts
@@ -1,22 +1,26 @@
-import { $ } from 'execa'
 import { buildSUUID } from '@peertube/peertube-node-utils'
-import { lstat } from 'node:fs/promises'
-import { OpenaiTranscriber, WhisperTranscribeArgs } from './openai-transcriber.js'
-import { TranscriptFile } from '../../transcript/index.js'
-import { WhisperBuiltinModel } from '../whisper-builtin-model.js'
+import { $ } from 'execa'
 import assert from 'node:assert'
+import { lstat } from 'node:fs/promises'
+import { TranscribeArgs } from '../../abstract-transcriber.js'
+import { TranscriptFile } from '../../transcript-file.js'
+import { TranscriptionModel } from '../../transcription-model.js'
+import { WhisperBuiltinModel } from '../whisper-builtin-model.js'
+import { OpenaiTranscriber } from './openai-transcriber.js'
 
 export class Ctranslate2Transcriber extends OpenaiTranscriber {
+
   async transcribe ({
     mediaFilePath,
     model = new WhisperBuiltinModel('tiny'),
     language,
-    format = 'vtt',
+    format,
+    transcriptDirectory,
     runId = buildSUUID()
-  }: WhisperTranscribeArgs): Promise<TranscriptFile> {
+  }: TranscribeArgs): Promise<TranscriptFile> {
     this.assertLanguageDetectionAvailable(language)
 
-    const $$ = $({ verbose: process.env.NODE_ENV !== 'production' })
+    const $$ = $({ verbose: process.env.NODE_ENV !== 'production', env: this.getExecEnv() })
 
     if (model.path) {
       assert(await lstat(model.path).then(stats => stats.isDirectory()), 'Model path must be a path to a directory.')
@@ -27,7 +31,7 @@ export class Ctranslate2Transcriber extends OpenaiTranscriber {
 
     this.createRun(runId)
     this.startRun()
-    await $$`${this.engine.binary} ${[
+    await $$`${this.getEngineBinary()} ${[
       mediaFilePath,
       ...modelArgs,
       '--word_timestamps',
@@ -35,15 +39,25 @@ export class Ctranslate2Transcriber extends OpenaiTranscriber {
       '--output_format',
       'all',
       '--output_dir',
-      this.transcriptDirectory,
+      transcriptDirectory,
       ...languageArgs
     ]}`
     this.stopRun()
 
     return new TranscriptFile({
-      language: language || await this.getDetectedLanguage(mediaFilePath),
-      path: this.getTranscriptFilePath(mediaFilePath, format),
+      language: language || await this.getDetectedLanguage(transcriptDirectory, mediaFilePath),
+      path: this.getTranscriptFilePath(transcriptDirectory, mediaFilePath, format),
       format
     })
   }
+
+  supports (model: TranscriptionModel) {
+    return model.format === 'CTranslate2'
+  }
+
+  async install (directory: string) {
+    const $$ = $({ verbose: true })
+
+    await $$`pip3 install -U -t ${directory} whisper-ctranslate2==${this.engine.version}`
+  }
 }
diff --git a/packages/transcription/src/whisper/transcriber/index.ts b/packages/transcription/src/whisper/transcriber/index.ts
index 950c39b07..3b9a2a745 100644
--- a/packages/transcription/src/whisper/transcriber/index.ts
+++ b/packages/transcription/src/whisper/transcriber/index.ts
@@ -1,3 +1,2 @@
 export * from './ctranslate2-transcriber.js'
 export * from './openai-transcriber.js'
-export * from './timestamped-transcriber.js'
diff --git a/packages/transcription/src/whisper/transcriber/openai-transcriber.ts b/packages/transcription/src/whisper/transcriber/openai-transcriber.ts
index 5d9a7ce85..20f07b615 100644
--- a/packages/transcription/src/whisper/transcriber/openai-transcriber.ts
+++ b/packages/transcription/src/whisper/transcriber/openai-transcriber.ts
@@ -1,31 +1,30 @@
-import { join } from 'path'
-import { $ } from 'execa'
 import { buildSUUID } from '@peertube/peertube-node-utils'
-import { TranscriptFile, TranscriptFormat } from '../../transcript/index.js'
-import { AbstractTranscriber, TranscribeArgs } from '../../abstract-transcriber.js'
-import { WhisperBuiltinModel } from '../whisper-builtin-model.js'
-import { TranscriptionModel } from '../../transcription-model.js'
-import { readFile } from 'node:fs/promises'
+import { $ } from 'execa'
+import { readJSON } from 'fs-extra/esm'
 import { parse } from 'node:path'
-
-export type WhisperTranscribeArgs = Omit<TranscribeArgs, 'model'> & { model?: TranscriptionModel }
+import { join, resolve } from 'path'
+import { AbstractTranscriber, TranscribeArgs } from '../../abstract-transcriber.js'
+import { TranscriptFile, TranscriptFormat } from '../../transcript-file.js'
 
 export class OpenaiTranscriber extends AbstractTranscriber {
+
   async transcribe ({
     mediaFilePath,
-    model = new WhisperBuiltinModel('tiny'),
+    model,
     language,
-    format = 'vtt',
+    format,
+    transcriptDirectory,
     runId = buildSUUID()
-  }: WhisperTranscribeArgs): Promise<TranscriptFile> {
+  }: TranscribeArgs): Promise<TranscriptFile> {
     this.assertLanguageDetectionAvailable(language)
 
-    const $$ = $({ verbose: process.env.NODE_ENV !== 'production' })
+    const $$ = $({ verbose: process.env.NODE_ENV !== 'production', env: this.getExecEnv() })
     const languageArgs = language ? [ '--language', language ] : []
 
     this.createRun(runId)
     this.startRun()
-    await $$`${this.engine.binary} ${[
+
+    await $$`${this.getEngineBinary()} ${[
       mediaFilePath,
       '--word_timestamps',
       'True',
@@ -34,29 +33,45 @@ export class OpenaiTranscriber extends AbstractTranscriber {
       '--output_format',
       'all',
       '--output_dir',
-      this.transcriptDirectory,
+      transcriptDirectory,
       ...languageArgs
     ]}`
     this.stopRun()
 
     return new TranscriptFile({
-      language: language || await this.getDetectedLanguage(mediaFilePath),
-      path: this.getTranscriptFilePath(mediaFilePath, format),
+      language: language || await this.getDetectedLanguage(transcriptDirectory, mediaFilePath),
+      path: this.getTranscriptFilePath(transcriptDirectory, mediaFilePath, format),
       format
     })
   }
 
-  async getDetectedLanguage (mediaFilePath: string) {
-    const { language } = await this.readJsonTranscriptFile(mediaFilePath)
+  // ---------------------------------------------------------------------------
+
+  protected async getDetectedLanguage (transcriptDirectory: string, mediaFilePath: string) {
+    const { language } = await this.readJsonTranscriptFile(transcriptDirectory, mediaFilePath)
 
     return language
   }
 
-  async readJsonTranscriptFile (mediaFilePath: string) {
-    return JSON.parse(await readFile(this.getTranscriptFilePath(mediaFilePath, 'json'), 'utf8'))
+  protected async readJsonTranscriptFile (transcriptDirectory: string, mediaFilePath: string) {
+    return readJSON(this.getTranscriptFilePath(transcriptDirectory, mediaFilePath, 'json'), 'utf8')
   }
 
-  getTranscriptFilePath (mediaFilePath: string, format: TranscriptFormat) {
-    return join(this.transcriptDirectory, `${parse(mediaFilePath).name}.${format}`)
+  protected getTranscriptFilePath (transcriptDirectory: string, mediaFilePath: string, format: TranscriptFormat) {
+    return join(transcriptDirectory, `${parse(mediaFilePath).name}.${format}`)
+  }
+
+  // ---------------------------------------------------------------------------
+
+  async install (directory: string) {
+    const $$ = $({ verbose: true })
+
+    await $$`pip3 install -U -t ${[ directory ]} openai-whisper==${this.engine.version}`
+  }
+
+  protected getExecEnv () {
+    if (!this.binDirectory) return undefined
+
+    return { PYTHONPATH: resolve(this.binDirectory, '../') }
   }
 }
diff --git a/packages/transcription/src/whisper/transcriber/timestamped-transcriber.ts b/packages/transcription/src/whisper/transcriber/timestamped-transcriber.ts
deleted file mode 100644
index fcdd33eb1..000000000
--- a/packages/transcription/src/whisper/transcriber/timestamped-transcriber.ts
+++ /dev/null
@@ -1,55 +0,0 @@
-import { $ } from 'execa'
-import { buildSUUID } from '@peertube/peertube-node-utils'
-import assert from 'node:assert'
-import { join, parse } from 'node:path'
-import { existsSync } from 'node:fs'
-import { rename } from 'node:fs/promises'
-import { TranscriptFile, TranscriptFormat } from '../../transcript/index.js'
-import { OpenaiTranscriber, WhisperTranscribeArgs } from './openai-transcriber.js'
-import { WhisperBuiltinModel } from '../whisper-builtin-model.js'
-
-export class WhisperTimestampedTranscriber extends OpenaiTranscriber {
-  async transcribe ({
-    mediaFilePath,
-    model = new WhisperBuiltinModel('tiny'),
-    language,
-    format = 'vtt',
-    runId = buildSUUID()
-  }: WhisperTranscribeArgs): Promise<TranscriptFile> {
-    this.assertLanguageDetectionAvailable(language)
-
-    const $$ = $({ verbose: process.env.NODE_ENV !== 'production' })
-    const languageArgs = language ? [ '--language', language ] : []
-
-    this.createRun(runId)
-    this.startRun()
-    await $$`${this.engine.binary} ${[
-      mediaFilePath,
-      '--model',
-      model?.path || model.name,
-      '--output_format',
-      'all',
-      '--output_dir',
-      this.transcriptDirectory,
-      ...languageArgs
-    ]}`
-    this.stopRun()
-
-    const internalTranscriptPath = this.getTranscriptFilePath(mediaFilePath, format, false)
-    const transcriptPath = join(this.transcriptDirectory, `${parse(mediaFilePath).name}.${format}`)
-    // Whisper timestamped output files with the video file extension by defaults, ex: video.mp4.vtt
-    // @see https://github.com/linto-ai/whisper-timestamped/issues/189
-    assert(existsSync(internalTranscriptPath), `${internalTranscriptPath} file doesn't exist.`)
-    await rename(internalTranscriptPath, transcriptPath)
-    // communiquer-lors-dune-classe-transplantee.mp4.words.json
-    return new TranscriptFile({
-      language: language || await this.getDetectedLanguage(mediaFilePath),
-      path: transcriptPath,
-      format
-    })
-  }
-
-  getTranscriptFilePath (mediaFilePath: string, format: TranscriptFormat, words = true) {
-    return join(this.transcriptDirectory, `${parse(mediaFilePath).base}${words ? '.words' : ''}.${format}`)
-  }
-}
diff --git a/packages/transcription/tsconfig.json b/packages/transcription/tsconfig.json
index 94971d65a..c390b2ae9 100644
--- a/packages/transcription/tsconfig.json
+++ b/packages/transcription/tsconfig.json
@@ -9,7 +9,6 @@
     { "path": "../models" },
     { "path": "../core-utils" },
     { "path": "../node-utils" },
-    { "path": "../jiwer" },
     { "path": "../server-commands" }
   ]
 }
diff --git a/scripts/ci.sh b/scripts/ci.sh
index 8d483b42c..88ae5d689 100755
--- a/scripts/ci.sh
+++ b/scripts/ci.sh
@@ -131,6 +131,8 @@ elif [ "$1" = "api-5" ]; then
 
     MOCHA_PARALLEL=true runJSTest "$1" $((2*$speedFactor)) $transcodingFiles $runnersFiles
 elif [ "$1" = "external-plugins" ]; then
+    npm run install-dependencies:transcription --workspace=@peertube/tests
+
     npm run build:server
     npm run build:tests
     npm run build:peertube-runner
@@ -147,12 +149,13 @@ elif [ "$1" = "lint" ]; then
 
     ( cd client && npm run lint )
 elif [ "$1" = "transcription" ]; then
-    npm run preinstall --workspace=@peertube/peertube-transcription --workspace=@peertube/peertube-jiwer
+    npm run install-dependencies:transcription --workspace=@peertube/tests
+
     npm run build:server
     npm run build:tests
 
     transcriptionFiles=$(findTestFiles ./packages/tests/dist/transcription)
-    jiwerFiles=$(findTestFiles ./packages/tests/dist/jiwer)
+    transcriptionDevToolsFiles=$(findTestFiles ./packages/tests/dist/transcription-devtools)
 
-    MOCHA_PARALLEL=true runJSTest "$1" $((3*$speedFactor)) $transcriptionFiles $jiwerFiles
+    MOCHA_PARALLEL=true runJSTest "$1" $((3*$speedFactor)) $transcriptionFiles $transcriptionDevToolsFiles
 fi
diff --git a/server/core/controllers/api/config.ts b/server/core/controllers/api/config.ts
index 9f44e4a89..7937104e6 100644
--- a/server/core/controllers/api/config.ts
+++ b/server/core/controllers/api/config.ts
@@ -1,10 +1,16 @@
+import { About, ActorImageType, ActorImageType_Type, CustomConfig, HttpStatusCode, UserRight } from '@peertube/peertube-models'
+import { createReqFiles } from '@server/helpers/express-utils.js'
+import { MIMETYPES } from '@server/initializers/constants.js'
+import { deleteLocalActorImageFile, updateLocalActorImageFiles } from '@server/lib/local-actor.js'
+import { ServerConfigManager } from '@server/lib/server-config-manager.js'
+import { ActorImageModel } from '@server/models/actor/actor-image.js'
+import { getServerActor } from '@server/models/application/application.js'
+import { ModelCache } from '@server/models/shared/model-cache.js'
 import express from 'express'
 import { remove, writeJSON } from 'fs-extra/esm'
 import snakeCase from 'lodash-es/snakeCase.js'
 import validator from 'validator'
-import { ServerConfigManager } from '@server/lib/server-config-manager.js'
-import { About, ActorImageType, ActorImageType_Type, CustomConfig, HttpStatusCode, UserRight } from '@peertube/peertube-models'
-import { auditLoggerFactory, CustomConfigAuditView, getAuditIdFromRes } from '../../helpers/audit-logger.js'
+import { CustomConfigAuditView, auditLoggerFactory, getAuditIdFromRes } from '../../helpers/audit-logger.js'
 import { objectConverter } from '../../helpers/core-utils.js'
 import { CONFIG, reloadConfig } from '../../initializers/config.js'
 import { ClientHtml } from '../../lib/html/client-html.js'
@@ -18,12 +24,6 @@ import {
   updateBannerValidator
 } from '../../middlewares/index.js'
 import { customConfigUpdateValidator, ensureConfigIsEditable } from '../../middlewares/validators/config.js'
-import { createReqFiles } from '@server/helpers/express-utils.js'
-import { MIMETYPES } from '@server/initializers/constants.js'
-import { deleteLocalActorImageFile, updateLocalActorImageFiles } from '@server/lib/local-actor.js'
-import { getServerActor } from '@server/models/application/application.js'
-import { ActorImageModel } from '@server/models/actor/actor-image.js'
-import { ModelCache } from '@server/models/shared/model-cache.js'
 
 const configRouter = express.Router()
 
@@ -385,6 +385,12 @@ function customConfig (): CustomConfig {
         enabled: CONFIG.VIDEO_STUDIO.REMOTE_RUNNERS.ENABLED
       }
     },
+    videoTranscription: {
+      enabled: CONFIG.VIDEO_TRANSCRIPTION.ENABLED,
+      remoteRunners: {
+        enabled: CONFIG.VIDEO_TRANSCRIPTION.REMOTE_RUNNERS.ENABLED
+      }
+    },
     videoFile: {
       update: {
         enabled: CONFIG.VIDEO_FILE.UPDATE.ENABLED
diff --git a/server/core/controllers/api/runners/jobs.ts b/server/core/controllers/api/runners/jobs.ts
index c7aada7c2..5b81062d8 100644
--- a/server/core/controllers/api/runners/jobs.ts
+++ b/server/core/controllers/api/runners/jobs.ts
@@ -1,4 +1,25 @@
-import express, { UploadFiles } from 'express'
+import {
+  AbortRunnerJobBody,
+  AcceptRunnerJobResult,
+  ErrorRunnerJobBody,
+  HttpStatusCode,
+  ListRunnerJobsQuery,
+  LiveRTMPHLSTranscodingUpdatePayload,
+  RequestRunnerJobResult,
+  RunnerJobState,
+  RunnerJobSuccessBody,
+  RunnerJobSuccessPayload,
+  RunnerJobType,
+  RunnerJobUpdateBody,
+  RunnerJobUpdatePayload,
+  ServerErrorCode,
+  TranscriptionSuccess,
+  UserRight,
+  VODAudioMergeTranscodingSuccess,
+  VODHLSTranscodingSuccess,
+  VODWebVideoTranscodingSuccess,
+  VideoStudioTranscodingSuccess
+} from '@peertube/peertube-models'
 import { retryTransactionWrapper } from '@server/helpers/database-utils.js'
 import { createReqFiles } from '@server/helpers/express-utils.js'
 import { logger, loggerTagsFactory } from '@server/helpers/logger.js'
@@ -28,33 +49,13 @@ import {
   successRunnerJobValidator,
   updateRunnerJobValidator
 } from '@server/middlewares/validators/runners/index.js'
-import { RunnerModel } from '@server/models/runner/runner.js'
 import { RunnerJobModel } from '@server/models/runner/runner-job.js'
-import {
-  AbortRunnerJobBody,
-  AcceptRunnerJobResult,
-  ErrorRunnerJobBody,
-  HttpStatusCode,
-  ListRunnerJobsQuery,
-  LiveRTMPHLSTranscodingUpdatePayload,
-  RequestRunnerJobResult,
-  RunnerJobState,
-  RunnerJobSuccessBody,
-  RunnerJobSuccessPayload,
-  RunnerJobType,
-  RunnerJobUpdateBody,
-  RunnerJobUpdatePayload,
-  ServerErrorCode,
-  UserRight,
-  VideoStudioTranscodingSuccess,
-  VODAudioMergeTranscodingSuccess,
-  VODHLSTranscodingSuccess,
-  VODWebVideoTranscodingSuccess
-} from '@peertube/peertube-models'
+import { RunnerModel } from '@server/models/runner/runner.js'
+import express, { UploadFiles } from 'express'
 
 const postRunnerJobSuccessVideoFiles = createReqFiles(
-  [ 'payload[videoFile]', 'payload[resolutionPlaylistFile]' ],
-  { ...MIMETYPES.VIDEO.MIMETYPE_EXT, ...MIMETYPES.M3U8.MIMETYPE_EXT }
+  [ 'payload[videoFile]', 'payload[resolutionPlaylistFile]', 'payload[vttFile]' ],
+  { ...MIMETYPES.VIDEO.MIMETYPE_EXT, ...MIMETYPES.M3U8.MIMETYPE_EXT, ...MIMETYPES.VIDEO_CAPTIONS.MIMETYPE_EXT }
 )
 
 const runnerJobUpdateVideoFiles = createReqFiles(
@@ -345,7 +346,15 @@ const jobSuccessPayloadBuilders: {
     }
   },
 
-  'live-rtmp-hls-transcoding': () => ({})
+  'live-rtmp-hls-transcoding': () => ({}),
+
+  'video-transcription': (payload: TranscriptionSuccess, files) => {
+    return {
+      ...payload,
+
+      vttFile: files['payload[vttFile]'][0].path
+    }
+  }
 }
 
 async function postRunnerJobSuccess (req: express.Request, res: express.Response) {
diff --git a/server/core/controllers/api/users/my-notifications.ts b/server/core/controllers/api/users/my-notifications.ts
index 3266dd811..cd37eeb00 100644
--- a/server/core/controllers/api/users/my-notifications.ts
+++ b/server/core/controllers/api/users/my-notifications.ts
@@ -76,6 +76,7 @@ async function updateNotificationSettings (req: express.Request, res: express.Re
     abuseStateChange: body.abuseStateChange,
     newPeerTubeVersion: body.newPeerTubeVersion,
     newPluginVersion: body.newPluginVersion,
+    myVideoTranscriptionGenerated: body.myVideoTranscriptionGenerated,
     myVideoStudioEditionFinished: body.myVideoStudioEditionFinished
   }
 
diff --git a/server/core/controllers/api/videos/captions.ts b/server/core/controllers/api/videos/captions.ts
index 9c3db2f90..7703eed0f 100644
--- a/server/core/controllers/api/videos/captions.ts
+++ b/server/core/controllers/api/videos/captions.ts
@@ -1,31 +1,47 @@
-import express from 'express'
-import { HttpStatusCode } from '@peertube/peertube-models'
+import { HttpStatusCode, VideoCaptionGenerate } from '@peertube/peertube-models'
 import { Hooks } from '@server/lib/plugins/hooks.js'
+import { createLocalCaption, createTranscriptionTaskIfNeeded } from '@server/lib/video-captions.js'
+import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
+import express from 'express'
 import { createReqFiles } from '../../../helpers/express-utils.js'
-import { logger } from '../../../helpers/logger.js'
+import { logger, loggerTagsFactory } from '../../../helpers/logger.js'
 import { getFormattedObjects } from '../../../helpers/utils.js'
 import { MIMETYPES } from '../../../initializers/constants.js'
 import { sequelizeTypescript } from '../../../initializers/database.js'
 import { federateVideoIfNeeded } from '../../../lib/activitypub/videos/index.js'
 import { asyncMiddleware, asyncRetryTransactionMiddleware, authenticate } from '../../../middlewares/index.js'
-import { addVideoCaptionValidator, deleteVideoCaptionValidator, listVideoCaptionsValidator } from '../../../middlewares/validators/index.js'
+import {
+  addVideoCaptionValidator,
+  deleteVideoCaptionValidator,
+  generateVideoCaptionValidator,
+  listVideoCaptionsValidator
+} from '../../../middlewares/validators/index.js'
 import { VideoCaptionModel } from '../../../models/video/video-caption.js'
-import { createLocalCaption } from '@server/lib/video-captions.js'
+
+const lTags = loggerTagsFactory('api', 'video-caption')
 
 const reqVideoCaptionAdd = createReqFiles([ 'captionfile' ], MIMETYPES.VIDEO_CAPTIONS.MIMETYPE_EXT)
 
 const videoCaptionsRouter = express.Router()
 
+videoCaptionsRouter.post('/:videoId/captions/generate',
+  authenticate,
+  asyncMiddleware(generateVideoCaptionValidator),
+  asyncMiddleware(createGenerateVideoCaption)
+)
+
 videoCaptionsRouter.get('/:videoId/captions',
   asyncMiddleware(listVideoCaptionsValidator),
   asyncMiddleware(listVideoCaptions)
 )
+
 videoCaptionsRouter.put('/:videoId/captions/:captionLanguage',
   authenticate,
   reqVideoCaptionAdd,
   asyncMiddleware(addVideoCaptionValidator),
   asyncRetryTransactionMiddleware(createVideoCaption)
 )
+
 videoCaptionsRouter.delete('/:videoId/captions/:captionLanguage',
   authenticate,
   asyncMiddleware(deleteVideoCaptionValidator),
@@ -40,6 +56,19 @@ export {
 
 // ---------------------------------------------------------------------------
 
+async function createGenerateVideoCaption (req: express.Request, res: express.Response) {
+  const video = res.locals.videoAll
+
+  const body = req.body as VideoCaptionGenerate
+  if (body.forceTranscription === true) {
+    await VideoJobInfoModel.abortAllTasks(video.uuid, 'pendingTranscription')
+  }
+
+  await createTranscriptionTaskIfNeeded(video)
+
+  return res.sendStatus(HttpStatusCode.NO_CONTENT_204)
+}
+
 async function listVideoCaptions (req: express.Request, res: express.Response) {
   const data = await VideoCaptionModel.listVideoCaptions(res.locals.onlyVideo.id)
 
@@ -74,7 +103,7 @@ async function deleteVideoCaption (req: express.Request, res: express.Response)
     await federateVideoIfNeeded(video, false, t)
   })
 
-  logger.info('Video caption %s of video %s deleted.', videoCaption.language, video.uuid)
+  logger.info('Video caption %s of video %s deleted.', videoCaption.language, video.uuid, lTags(video.uuid))
 
   Hooks.runAction('action:api.video-caption.deleted', { caption: videoCaption, req, res })
 
diff --git a/server/core/controllers/api/videos/import.ts b/server/core/controllers/api/videos/import.ts
index 5e6694899..69dc368bd 100644
--- a/server/core/controllers/api/videos/import.ts
+++ b/server/core/controllers/api/videos/import.ts
@@ -142,7 +142,8 @@ async function handleTorrentImport (req: express.Request, res: express.Response,
       ? 'torrent-file'
       : 'magnet-uri',
     videoImportId: videoImport.id,
-    preventException: false
+    preventException: false,
+    generateTranscription: body.generateTranscription
   }
   await JobQueue.Instance.createJob({ type: 'video-import', payload })
 
diff --git a/server/core/helpers/custom-validators/runners/jobs.ts b/server/core/helpers/custom-validators/runners/jobs.ts
index 78c6c2e1e..ec0238b77 100644
--- a/server/core/helpers/custom-validators/runners/jobs.ts
+++ b/server/core/helpers/custom-validators/runners/jobs.ts
@@ -1,83 +1,74 @@
-import { UploadFilesForCheck } from 'express'
-import validator from 'validator'
-import { CONSTRAINTS_FIELDS, RUNNER_JOB_STATES } from '@server/initializers/constants.js'
 import {
   LiveRTMPHLSTranscodingSuccess,
   RunnerJobSuccessPayload,
   RunnerJobType,
   RunnerJobUpdatePayload,
-  VideoStudioTranscodingSuccess,
+  TranscriptionSuccess,
   VODAudioMergeTranscodingSuccess,
   VODHLSTranscodingSuccess,
-  VODWebVideoTranscodingSuccess
+  VODWebVideoTranscodingSuccess,
+  VideoStudioTranscodingSuccess
 } from '@peertube/peertube-models'
+import { CONSTRAINTS_FIELDS, RUNNER_JOB_STATES } from '@server/initializers/constants.js'
+import { UploadFilesForCheck } from 'express'
+import validator from 'validator'
 import { exists, isArray, isFileValid, isSafeFilename } from '../misc.js'
 
 const RUNNER_JOBS_CONSTRAINTS_FIELDS = CONSTRAINTS_FIELDS.RUNNER_JOBS
 
 const runnerJobTypes = new Set([ 'vod-hls-transcoding', 'vod-web-video-transcoding', 'vod-audio-merge-transcoding' ])
-function isRunnerJobTypeValid (value: RunnerJobType) {
+export function isRunnerJobTypeValid (value: RunnerJobType) {
   return runnerJobTypes.has(value)
 }
 
-function isRunnerJobSuccessPayloadValid (value: RunnerJobSuccessPayload, type: RunnerJobType, files: UploadFilesForCheck) {
+export function isRunnerJobSuccessPayloadValid (value: RunnerJobSuccessPayload, type: RunnerJobType, files: UploadFilesForCheck) {
   return isRunnerJobVODWebVideoResultPayloadValid(value as VODWebVideoTranscodingSuccess, type, files) ||
     isRunnerJobVODHLSResultPayloadValid(value as VODHLSTranscodingSuccess, type, files) ||
     isRunnerJobVODAudioMergeResultPayloadValid(value as VODHLSTranscodingSuccess, type, files) ||
     isRunnerJobLiveRTMPHLSResultPayloadValid(value as LiveRTMPHLSTranscodingSuccess, type) ||
-    isRunnerJobVideoStudioResultPayloadValid(value as VideoStudioTranscodingSuccess, type, files)
+    isRunnerJobVideoStudioResultPayloadValid(value as VideoStudioTranscodingSuccess, type, files) ||
+    isRunnerJobTranscriptionResultPayloadValid(value as TranscriptionSuccess, type, files)
 }
 
 // ---------------------------------------------------------------------------
 
-function isRunnerJobProgressValid (value: string) {
+export function isRunnerJobProgressValid (value: string) {
   return validator.default.isInt(value + '', RUNNER_JOBS_CONSTRAINTS_FIELDS.PROGRESS)
 }
 
-function isRunnerJobUpdatePayloadValid (value: RunnerJobUpdatePayload, type: RunnerJobType, files: UploadFilesForCheck) {
+export function isRunnerJobUpdatePayloadValid (value: RunnerJobUpdatePayload, type: RunnerJobType, files: UploadFilesForCheck) {
   return isRunnerJobVODWebVideoUpdatePayloadValid(value, type, files) ||
     isRunnerJobVODHLSUpdatePayloadValid(value, type, files) ||
     isRunnerJobVideoStudioUpdatePayloadValid(value, type, files) ||
     isRunnerJobVODAudioMergeUpdatePayloadValid(value, type, files) ||
-    isRunnerJobLiveRTMPHLSUpdatePayloadValid(value, type, files)
+    isRunnerJobLiveRTMPHLSUpdatePayloadValid(value, type, files) ||
+    isRunnerJobTranscriptionUpdatePayloadValid(value, type, files)
 }
 
 // ---------------------------------------------------------------------------
 
-function isRunnerJobTokenValid (value: string) {
+export function isRunnerJobTokenValid (value: string) {
   return exists(value) && validator.default.isLength(value, RUNNER_JOBS_CONSTRAINTS_FIELDS.TOKEN)
 }
 
-function isRunnerJobAbortReasonValid (value: string) {
+export function isRunnerJobAbortReasonValid (value: string) {
   return validator.default.isLength(value, RUNNER_JOBS_CONSTRAINTS_FIELDS.REASON)
 }
 
-function isRunnerJobErrorMessageValid (value: string) {
+export function isRunnerJobErrorMessageValid (value: string) {
   return validator.default.isLength(value, RUNNER_JOBS_CONSTRAINTS_FIELDS.ERROR_MESSAGE)
 }
 
-function isRunnerJobStateValid (value: any) {
+export function isRunnerJobStateValid (value: any) {
   return exists(value) && RUNNER_JOB_STATES[value] !== undefined
 }
 
-function isRunnerJobArrayOfStateValid (value: any) {
+export function isRunnerJobArrayOfStateValid (value: any) {
   return isArray(value) && value.every(v => isRunnerJobStateValid(v))
 }
 
 // ---------------------------------------------------------------------------
-
-export {
-  isRunnerJobTypeValid,
-  isRunnerJobSuccessPayloadValid,
-  isRunnerJobUpdatePayloadValid,
-  isRunnerJobTokenValid,
-  isRunnerJobErrorMessageValid,
-  isRunnerJobProgressValid,
-  isRunnerJobAbortReasonValid,
-  isRunnerJobArrayOfStateValid,
-  isRunnerJobStateValid
-}
-
+// Private
 // ---------------------------------------------------------------------------
 
 function isRunnerJobVODWebVideoResultPayloadValid (
@@ -124,6 +115,15 @@ function isRunnerJobVideoStudioResultPayloadValid (
     isFileValid({ files, field: 'payload[videoFile]', mimeTypeRegex: null, maxSize: null })
 }
 
+function isRunnerJobTranscriptionResultPayloadValid (
+  value: TranscriptionSuccess,
+  type: RunnerJobType,
+  files: UploadFilesForCheck
+) {
+  return type === 'video-transcription' &&
+    isFileValid({ files, field: 'payload[vttFile]', mimeTypeRegex: null, maxSize: null })
+}
+
 // ---------------------------------------------------------------------------
 
 function isRunnerJobVODWebVideoUpdatePayloadValid (
@@ -153,6 +153,15 @@ function isRunnerJobVODAudioMergeUpdatePayloadValid (
     (!value || (typeof value === 'object' && Object.keys(value).length === 0))
 }
 
+function isRunnerJobTranscriptionUpdatePayloadValid (
+  value: RunnerJobUpdatePayload,
+  type: RunnerJobType,
+  _files: UploadFilesForCheck
+) {
+  return type === 'video-transcription' &&
+    (!value || (typeof value === 'object' && Object.keys(value).length === 0))
+}
+
 function isRunnerJobLiveRTMPHLSUpdatePayloadValid (
   value: RunnerJobUpdatePayload,
   type: RunnerJobType,
diff --git a/server/core/initializers/config.ts b/server/core/initializers/config.ts
index 5fe0f8376..2ecffcb36 100644
--- a/server/core/initializers/config.ts
+++ b/server/core/initializers/config.ts
@@ -13,6 +13,7 @@ import {
 import { decacheModule } from '@server/helpers/decache.js'
 import { buildPath, root } from '@peertube/peertube-node-utils'
 import { parseBytes, parseDurationToMs } from '../helpers/core-utils.js'
+import { TranscriptionEngineName, WhisperBuiltinModelName } from '@peertube/peertube-transcription'
 
 const require = createRequire(import.meta.url)
 let config: IConfig = require('config')
@@ -515,6 +516,16 @@ const CONFIG = {
       get ENABLED () { return config.get<boolean>('video_file.update.enabled') }
     }
   },
+  VIDEO_TRANSCRIPTION: {
+    get ENABLED () { return config.get<boolean>('video_transcription.enabled') },
+    get ENGINE () { return config.get<TranscriptionEngineName>('video_transcription.engine') },
+    get ENGINE_PATH () { return config.get<string>('video_transcription.engine_path') },
+    get MODEL () { return config.get<WhisperBuiltinModelName>('video_transcription.model') },
+    get MODEL_PATH () { return config.get<string>('video_transcription.model_path') },
+    REMOTE_RUNNERS: {
+      get ENABLED () { return config.get<boolean>('video_transcription.remote_runners.enabled') }
+    }
+  },
   IMPORT: {
     VIDEOS: {
       get CONCURRENCY () { return config.get<number>('import.videos.concurrency') },
diff --git a/server/core/initializers/constants.ts b/server/core/initializers/constants.ts
index ebb205a33..faf944249 100644
--- a/server/core/initializers/constants.ts
+++ b/server/core/initializers/constants.ts
@@ -47,7 +47,7 @@ import { cpus } from 'os'
 
 // ---------------------------------------------------------------------------
 
-const LAST_MIGRATION_VERSION = 850
+const LAST_MIGRATION_VERSION = 855
 
 // ---------------------------------------------------------------------------
 
@@ -213,7 +213,8 @@ const JOB_ATTEMPTS: { [id in JobType]: number } = {
   'notify': 1,
   'federate-video': 1,
   'create-user-export': 1,
-  'import-user-archive': 1
+  'import-user-archive': 1,
+  'video-transcription': 1
 }
 // Excluded keys are jobs that can be configured by admins
 const JOB_CONCURRENCY: { [id in Exclude<JobType, 'video-transcoding' | 'video-import'>]: number } = {
@@ -241,7 +242,8 @@ const JOB_CONCURRENCY: { [id in Exclude<JobType, 'video-transcoding' | 'video-im
   'notify': 5,
   'federate-video': 3,
   'create-user-export': 1,
-  'import-user-archive': 1
+  'import-user-archive': 1,
+  'video-transcription': 1
 }
 const JOB_TTL: { [id in JobType]: number } = {
   'activitypub-http-broadcast': 60000 * 10, // 10 minutes
@@ -270,7 +272,8 @@ const JOB_TTL: { [id in JobType]: number } = {
   'notify': 60000 * 5, // 5 minutes
   'federate-video': 60000 * 5, // 5 minutes,
   'create-user-export': 60000 * 60 * 24, // 24 hours
-  'import-user-archive': 60000 * 60 * 24 // 24 hours
+  'import-user-archive': 60000 * 60 * 24, // 24 hours
+  'video-transcription': 1000 * 3600 * 6 // 6 hours
 }
 const REPEAT_JOBS: { [ id in JobType ]?: RepeatOptions } = {
   'videos-views-stats': {
@@ -282,7 +285,8 @@ const REPEAT_JOBS: { [ id in JobType ]?: RepeatOptions } = {
 }
 const JOB_PRIORITY = {
   TRANSCODING: 100,
-  VIDEO_STUDIO: 150
+  VIDEO_STUDIO: 150,
+  TRANSCRIPTION: 200
 }
 
 const JOB_REMOVAL_OPTIONS = {
@@ -1013,7 +1017,9 @@ const DIRECTORIES = {
 
   ORIGINAL_VIDEOS: CONFIG.STORAGE.ORIGINAL_VIDEO_FILES_DIR,
 
-  HLS_REDUNDANCY: join(CONFIG.STORAGE.REDUNDANCY_DIR, 'hls')
+  HLS_REDUNDANCY: join(CONFIG.STORAGE.REDUNDANCY_DIR, 'hls'),
+
+  LOCAL_PIP_DIRECTORY: join(CONFIG.STORAGE.BIN_DIR, 'pip')
 }
 
 const RESUMABLE_UPLOAD_SESSION_LIFETIME = SCHEDULER_INTERVALS_MS.REMOVE_DANGLING_RESUMABLE_UPLOADS
diff --git a/server/core/initializers/migrations/0855-transcription.ts b/server/core/initializers/migrations/0855-transcription.ts
new file mode 100644
index 000000000..330e67e70
--- /dev/null
+++ b/server/core/initializers/migrations/0855-transcription.ts
@@ -0,0 +1,67 @@
+import * as Sequelize from 'sequelize'
+
+async function up (utils: {
+  transaction: Sequelize.Transaction
+  queryInterface: Sequelize.QueryInterface
+  sequelize: Sequelize.Sequelize
+}): Promise<void> {
+  const { transaction } = utils
+
+  // Notification
+  {
+    await utils.queryInterface.addColumn('userNotification', 'videoCaptionId', {
+      type: Sequelize.INTEGER,
+      defaultValue: null,
+      allowNull: true,
+      references: {
+        model: 'videoCaption',
+        key: 'id'
+      },
+      onUpdate: 'CASCADE',
+      onDelete: 'CASCADE'
+    }, { transaction })
+  }
+
+  // Notification settings
+  {
+    {
+      const data = {
+        type: Sequelize.INTEGER,
+        defaultValue: null,
+        allowNull: true
+      }
+      await utils.queryInterface.addColumn('userNotificationSetting', 'myVideoTranscriptionGenerated', data, { transaction })
+    }
+
+    {
+      const query = 'UPDATE "userNotificationSetting" SET "myVideoTranscriptionGenerated" = 1'
+      await utils.sequelize.query(query, { transaction })
+    }
+
+    {
+      const data = {
+        type: Sequelize.INTEGER,
+        defaultValue: null,
+        allowNull: false
+      }
+      await utils.queryInterface.changeColumn('userNotificationSetting', 'myVideoTranscriptionGenerated', data, { transaction })
+    }
+  }
+
+  // Video job info
+  {
+    await utils.queryInterface.addColumn('videoJobInfo', 'pendingTranscription', {
+      type: Sequelize.INTEGER,
+      allowNull: false,
+      defaultValue: 0
+    }, { transaction })
+  }
+}
+
+function down (options) {
+  throw new Error('Not implemented.')
+}
+
+export {
+  down, up
+}
diff --git a/server/core/lib/job-queue/handlers/video-import.ts b/server/core/lib/job-queue/handlers/video-import.ts
index c791a402b..14337a85e 100644
--- a/server/core/lib/job-queue/handlers/video-import.ts
+++ b/server/core/lib/job-queue/handlers/video-import.ts
@@ -24,11 +24,13 @@ import { Hooks } from '@server/lib/plugins/hooks.js'
 import { ServerConfigManager } from '@server/lib/server-config-manager.js'
 import { createOptimizeOrMergeAudioJobs } from '@server/lib/transcoding/create-transcoding-job.js'
 import { isUserQuotaValid } from '@server/lib/user.js'
+import { createTranscriptionTaskIfNeeded } from '@server/lib/video-captions.js'
 import { replaceChaptersIfNotExist } from '@server/lib/video-chapters.js'
 import { buildNewFile } from '@server/lib/video-file.js'
 import { buildMoveJob, buildStoryboardJobIfNeeded } from '@server/lib/video-jobs.js'
 import { VideoPathManager } from '@server/lib/video-path-manager.js'
 import { buildNextVideoState } from '@server/lib/video-state.js'
+import { VideoCaptionModel } from '@server/models/video/video-caption.js'
 import { MUserId, MVideoFile, MVideoFullLight } from '@server/types/models/index.js'
 import { MVideoImport, MVideoImportDefault, MVideoImportDefaultFiles, MVideoImportVideo } from '@server/types/models/video/video-import.js'
 import { Job } from 'bullmq'
@@ -84,7 +86,7 @@ export {
 async function processTorrentImport (job: Job, videoImport: MVideoImportDefault, payload: VideoImportTorrentPayload) {
   logger.info('Processing torrent video import in job %s.', job.id)
 
-  const options = { type: payload.type, videoImportId: payload.videoImportId }
+  const options = { type: payload.type, generateTranscription: payload.generateTranscription, videoImportId: payload.videoImportId }
 
   const target = {
     torrentName: videoImport.torrentName ? getSecureTorrentName(videoImport.torrentName) : undefined,
@@ -96,7 +98,7 @@ async function processTorrentImport (job: Job, videoImport: MVideoImportDefault,
 async function processYoutubeDLImport (job: Job, videoImport: MVideoImportDefault, payload: VideoImportYoutubeDLPayload) {
   logger.info('Processing youtubeDL video import in job %s.', job.id)
 
-  const options = { type: payload.type, videoImportId: videoImport.id }
+  const options = { type: payload.type, generateTranscription: payload.generateTranscription, videoImportId: videoImport.id }
 
   const youtubeDL = new YoutubeDLWrapper(
     videoImport.targetUrl,
@@ -122,6 +124,7 @@ async function getVideoImportOrDie (payload: VideoImportPayload) {
 
 type ProcessFileOptions = {
   type: VideoImportYoutubeDLPayloadType | VideoImportTorrentPayloadType
+  generateTranscription: boolean
   videoImportId: number
 }
 async function processFile (downloader: () => Promise<string>, videoImport: MVideoImportDefault, options: ProcessFileOptions) {
@@ -228,7 +231,14 @@ async function processFile (downloader: () => Promise<string>, videoImport: MVid
         })
       })
 
-      await afterImportSuccess({ videoImport: videoImportUpdated, video, videoFile, user: videoImport.User, videoFileAlreadyLocked: true })
+      await afterImportSuccess({
+        videoImport: videoImportUpdated,
+        video,
+        videoFile,
+        user: videoImport.User,
+        generateTranscription: options.generateTranscription,
+        videoFileAlreadyLocked: true
+      })
     } finally {
       videoFileLockReleaser()
     }
@@ -277,9 +287,12 @@ async function afterImportSuccess (options: {
   video: MVideoFullLight
   videoFile: MVideoFile
   user: MUserId
+
   videoFileAlreadyLocked: boolean
+
+  generateTranscription: boolean
 }) {
-  const { video, videoFile, videoImport, user, videoFileAlreadyLocked } = options
+  const { video, videoFile, videoImport, user, generateTranscription, videoFileAlreadyLocked } = options
 
   Notifier.Instance.notifyOnFinishedVideoImport({ videoImport: Object.assign(videoImport, { Video: video }), success: true })
 
@@ -294,6 +307,10 @@ async function afterImportSuccess (options: {
   // Generate the storyboard in the job queue, and don't forget to federate an update after
   await JobQueue.Instance.createJob(buildStoryboardJobIfNeeded({ video, federate: true }))
 
+  if (await VideoCaptionModel.hasVideoCaption(video.id) !== true && generateTranscription === true) {
+    await createTranscriptionTaskIfNeeded(video)
+  }
+
   if (video.state === VideoState.TO_MOVE_TO_EXTERNAL_STORAGE) {
     await JobQueue.Instance.createJob(
       await buildMoveJob({ video, previousVideoState: VideoState.TO_IMPORT, type: 'move-to-object-storage' })
diff --git a/server/core/lib/job-queue/handlers/video-live-ending.ts b/server/core/lib/job-queue/handlers/video-live-ending.ts
index 496367eda..91ba79a63 100644
--- a/server/core/lib/job-queue/handlers/video-live-ending.ts
+++ b/server/core/lib/job-queue/handlers/video-live-ending.ts
@@ -1,9 +1,7 @@
-import { Job } from 'bullmq'
-import { remove } from 'fs-extra/esm'
-import { readdir } from 'fs/promises'
-import { join } from 'path'
+import { ffprobePromise, getAudioStream, getVideoStreamDimensionsInfo, getVideoStreamFPS } from '@peertube/peertube-ffmpeg'
 import { ThumbnailType, VideoLiveEndingPayload, VideoState } from '@peertube/peertube-models'
 import { peertubeTruncate } from '@server/helpers/core-utils.js'
+import { CONFIG } from '@server/initializers/config.js'
 import { CONSTRAINTS_FIELDS } from '@server/initializers/constants.js'
 import { getLocalVideoActivityPubUrl } from '@server/lib/activitypub/url.js'
 import { federateVideoIfNeeded } from '@server/lib/activitypub/videos/index.js'
@@ -16,7 +14,10 @@ import {
 } from '@server/lib/paths.js'
 import { generateLocalVideoMiniature, regenerateMiniaturesIfNeeded } from '@server/lib/thumbnail.js'
 import { generateHlsPlaylistResolutionFromTS } from '@server/lib/transcoding/hls-transcoding.js'
+import { createTranscriptionTaskIfNeeded } from '@server/lib/video-captions.js'
+import { buildStoryboardJobIfNeeded } from '@server/lib/video-jobs.js'
 import { VideoPathManager } from '@server/lib/video-path-manager.js'
+import { isVideoInPublicDirectory } from '@server/lib/video-privacy.js'
 import { moveToNextState } from '@server/lib/video-state.js'
 import { VideoBlacklistModel } from '@server/models/video/video-blacklist.js'
 import { VideoFileModel } from '@server/models/video/video-file.js'
@@ -26,11 +27,12 @@ import { VideoLiveModel } from '@server/models/video/video-live.js'
 import { VideoStreamingPlaylistModel } from '@server/models/video/video-streaming-playlist.js'
 import { VideoModel } from '@server/models/video/video.js'
 import { MVideo, MVideoLive, MVideoLiveSession, MVideoWithAllFiles } from '@server/types/models/index.js'
-import { ffprobePromise, getAudioStream, getVideoStreamDimensionsInfo, getVideoStreamFPS } from '@peertube/peertube-ffmpeg'
+import { Job } from 'bullmq'
+import { remove } from 'fs-extra/esm'
+import { readdir } from 'fs/promises'
+import { join } from 'path'
 import { logger, loggerTagsFactory } from '../../../helpers/logger.js'
 import { JobQueue } from '../job-queue.js'
-import { isVideoInPublicDirectory } from '@server/lib/video-privacy.js'
-import { buildStoryboardJobIfNeeded } from '@server/lib/video-jobs.js'
 
 const lTags = loggerTagsFactory('live', 'job')
 
@@ -174,9 +176,13 @@ async function saveReplayToExternalVideo (options: {
     await replayVideo.addAndSaveThumbnail(thumbnail)
   }
 
-  await moveToNextState({ video: replayVideo, isNewVideo: true })
-
   await createStoryboardJob(replayVideo)
+
+  if (CONFIG.VIDEO_TRANSCRIPTION.ENABLED === true) {
+    await createTranscriptionTaskIfNeeded(replayVideo)
+  }
+
+  await moveToNextState({ video: replayVideo, isNewVideo: true })
 }
 
 async function replaceLiveByReplay (options: {
@@ -245,6 +251,10 @@ async function replaceLiveByReplay (options: {
   await moveToNextState({ video: videoWithFiles, isNewVideo: true })
 
   await createStoryboardJob(videoWithFiles)
+
+  if (CONFIG.VIDEO_TRANSCRIPTION.ENABLED === true) {
+    await createTranscriptionTaskIfNeeded(videoWithFiles)
+  }
 }
 
 async function assignReplayFilesToVideo (options: {
diff --git a/server/core/lib/job-queue/handlers/video-transcription.ts b/server/core/lib/job-queue/handlers/video-transcription.ts
new file mode 100644
index 000000000..47ae8e337
--- /dev/null
+++ b/server/core/lib/job-queue/handlers/video-transcription.ts
@@ -0,0 +1,21 @@
+import { VideoTranscriptionPayload } from '@peertube/peertube-models'
+import { generateSubtitle } from '@server/lib/video-captions.js'
+import { Job } from 'bullmq'
+import { logger, loggerTagsFactory } from '../../../helpers/logger.js'
+import { VideoModel } from '../../../models/video/video.js'
+
+const lTags = loggerTagsFactory('transcription')
+
+export async function processVideoTranscription (job: Job) {
+  const payload = job.data as VideoTranscriptionPayload
+
+  logger.info('Processing video transcription in job %s.', job.id)
+
+  const video = await VideoModel.load(payload.videoUUID)
+  if (!video) {
+    logger.info('Do not process transcription job %d, video does not exist.', job.id, lTags(payload.videoUUID))
+    return
+  }
+
+  return generateSubtitle({ video })
+}
diff --git a/server/core/lib/job-queue/job-queue.ts b/server/core/lib/job-queue/job-queue.ts
index 5559c8ea3..f3ecd952c 100644
--- a/server/core/lib/job-queue/job-queue.ts
+++ b/server/core/lib/job-queue/job-queue.ts
@@ -1,15 +1,3 @@
-import {
-  FlowJob,
-  FlowProducer,
-  Job,
-  JobsOptions,
-  Queue,
-  QueueEvents,
-  QueueEventsOptions,
-  QueueOptions,
-  Worker,
-  WorkerOptions
-} from 'bullmq'
 import { pick, timeoutPromise } from '@peertube/peertube-core-utils'
 import {
   ActivitypubFollowPayload,
@@ -37,12 +25,25 @@ import {
   VideoLiveEndingPayload,
   VideoRedundancyPayload,
   VideoStudioEditionPayload,
-  VideoTranscodingPayload
+  VideoTranscodingPayload,
+  VideoTranscriptionPayload
 } from '@peertube/peertube-models'
 import { parseDurationToMs } from '@server/helpers/core-utils.js'
 import { jobStates } from '@server/helpers/custom-validators/jobs.js'
 import { CONFIG } from '@server/initializers/config.js'
 import { processVideoRedundancy } from '@server/lib/job-queue/handlers/video-redundancy.js'
+import {
+  FlowJob,
+  FlowProducer,
+  Job,
+  JobsOptions,
+  Queue,
+  QueueEvents,
+  QueueEventsOptions,
+  QueueOptions,
+  Worker,
+  WorkerOptions
+} from 'bullmq'
 import { logger } from '../../helpers/logger.js'
 import { JOB_ATTEMPTS, JOB_CONCURRENCY, JOB_REMOVAL_OPTIONS, JOB_TTL, REPEAT_JOBS, WEBSERVER } from '../../initializers/constants.js'
 import { Hooks } from '../plugins/hooks.js'
@@ -58,10 +59,13 @@ import { processActivityPubHttpUnicast } from './handlers/activitypub-http-unica
 import { refreshAPObject } from './handlers/activitypub-refresher.js'
 import { processActorKeys } from './handlers/actor-keys.js'
 import { processAfterVideoChannelImport } from './handlers/after-video-channel-import.js'
+import { processCreateUserExport } from './handlers/create-user-export.js'
 import { processEmail } from './handlers/email.js'
 import { processFederateVideo } from './handlers/federate-video.js'
 import { processGenerateStoryboard } from './handlers/generate-storyboard.js'
+import { processImportUserArchive } from './handlers/import-user-archive.js'
 import { processManageVideoTorrent } from './handlers/manage-video-torrent.js'
+import { onMoveToFileSystemFailure, processMoveToFileSystem } from './handlers/move-to-file-system.js'
 import { onMoveToObjectStorageFailure, processMoveToObjectStorage } from './handlers/move-to-object-storage.js'
 import { processNotify } from './handlers/notify.js'
 import { processTranscodingJobBuilder } from './handlers/transcoding-job-builder.js'
@@ -71,10 +75,8 @@ import { processVideoImport } from './handlers/video-import.js'
 import { processVideoLiveEnding } from './handlers/video-live-ending.js'
 import { processVideoStudioEdition } from './handlers/video-studio-edition.js'
 import { processVideoTranscoding } from './handlers/video-transcoding.js'
+import { processVideoTranscription } from './handlers/video-transcription.js'
 import { processVideosViewsStats } from './handlers/video-views-stats.js'
-import { onMoveToFileSystemFailure, processMoveToFileSystem } from './handlers/move-to-file-system.js'
-import { processCreateUserExport } from './handlers/create-user-export.js'
-import { processImportUserArchive } from './handlers/import-user-archive.js'
 
 export type CreateJobArgument =
   { type: 'activitypub-http-broadcast', payload: ActivitypubHttpBroadcastPayload } |
@@ -104,7 +106,8 @@ export type CreateJobArgument =
   { type: 'federate-video', payload: FederateVideoPayload } |
   { type: 'create-user-export', payload: CreateUserExportPayload } |
   { type: 'generate-video-storyboard', payload: GenerateStoryboardPayload } |
-  { type: 'import-user-archive', payload: ImportUserArchivePayload }
+  { type: 'import-user-archive', payload: ImportUserArchivePayload } |
+  { type: 'video-transcription', payload: VideoTranscriptionPayload }
 
 export type CreateJobOptions = {
   delay?: number
@@ -139,7 +142,8 @@ const handlers: { [id in JobType]: (job: Job) => Promise<any> } = {
   'videos-views-stats': processVideosViewsStats,
   'generate-video-storyboard': processGenerateStoryboard,
   'create-user-export': processCreateUserExport,
-  'import-user-archive': processImportUserArchive
+  'import-user-archive': processImportUserArchive,
+  'video-transcription': processVideoTranscription
 }
 
 const errorHandlers: { [id in JobType]?: (job: Job, err: any) => Promise<any> } = {
@@ -171,10 +175,11 @@ const jobTypes: JobType[] = [
   'video-live-ending',
   'video-redundancy',
   'video-studio-edition',
-  'video-transcoding',
+  'video-transcription',
   'videos-views-stats',
   'create-user-export',
-  'import-user-archive'
+  'import-user-archive',
+  'video-transcoding'
 ]
 
 const silentFailure = new Set<JobType>([ 'activitypub-http-unicast' ])
@@ -561,6 +566,5 @@ class JobQueue {
 // ---------------------------------------------------------------------------
 
 export {
-  jobTypes,
-  JobQueue
+  JobQueue, jobTypes
 }
diff --git a/server/core/lib/local-video-creator.ts b/server/core/lib/local-video-creator.ts
index 4414a52a8..1c4c574d1 100644
--- a/server/core/lib/local-video-creator.ts
+++ b/server/core/lib/local-video-creator.ts
@@ -107,6 +107,8 @@ export class LocalVideoCreator {
     this.channel = options.channel
 
     this.videoAttributeResultHook = options.videoAttributeResultHook
+
+    this.lTags = options.lTags
   }
 
   async create () {
@@ -201,8 +203,11 @@ export class LocalVideoCreator {
 
         if (this.videoFile) {
           transaction.afterCommit(() => {
-            addVideoJobsAfterCreation({ video: this.video, videoFile: this.videoFile })
-              .catch(err => logger.error('Cannot build new video jobs of %s.', this.video.uuid, { err, ...this.lTags(this.video.uuid) }))
+            addVideoJobsAfterCreation({
+              video: this.video,
+              videoFile: this.videoFile,
+              generateTranscription: this.videoAttributes.generateTranscription ?? true
+            }).catch(err => logger.error('Cannot build new video jobs of %s.', this.video.uuid, { err, ...this.lTags(this.video.uuid) }))
           })
         } else {
           await federateVideoIfNeeded(this.video, true, transaction)
diff --git a/server/core/lib/notifier/notifier.ts b/server/core/lib/notifier/notifier.ts
index 277f0afb7..0048340d8 100644
--- a/server/core/lib/notifier/notifier.ts
+++ b/server/core/lib/notifier/notifier.ts
@@ -11,6 +11,7 @@ import {
   MCommentOwnerVideo,
   MPlugin,
   MVideoAccountLight,
+  MVideoCaptionVideo,
   MVideoFullLight
 } from '../../types/models/index.js'
 import { JobQueue } from '../job-queue/index.js'
@@ -41,7 +42,8 @@ import {
   OwnedPublicationAfterTranscoding,
   RegistrationRequestForModerators,
   StudioEditionFinishedForOwner,
-  UnblacklistForOwner
+  UnblacklistForOwner,
+  VideoTranscriptionGeneratedForOwner
 } from './shared/index.js'
 
 const lTags = loggerTagsFactory('notifier')
@@ -69,7 +71,8 @@ class Notifier {
     newAbuseMessage: [ NewAbuseMessageForReporter, NewAbuseMessageForModerators ],
     newPeertubeVersion: [ NewPeerTubeVersionForAdmins ],
     newPluginVersion: [ NewPluginVersionForAdmins ],
-    videoStudioEditionFinished: [ StudioEditionFinishedForOwner ]
+    videoStudioEditionFinished: [ StudioEditionFinishedForOwner ],
+    videoTranscriptionGenerated: [ VideoTranscriptionGeneratedForOwner ]
   }
 
   private static instance: Notifier
@@ -275,6 +278,16 @@ class Notifier {
       .catch(err => logger.error('Cannot notify on finished studio edition %s.', video.url, { err }))
   }
 
+  notifyOfGeneratedVideoTranscription (caption: MVideoCaptionVideo) {
+    const models = this.notificationModels.videoTranscriptionGenerated
+    const video = caption.Video
+
+    logger.debug('Notify on generated video transcription', { language: caption.language, video: video.url, ...lTags() })
+
+    this.sendNotifications(models, caption)
+      .catch(err => logger.error('Cannot notify on generated video transcription %s of video %s.', caption.language, video.url, { err }))
+  }
+
   private async notify <T> (object: AbstractNotification<T>) {
     await object.prepare()
 
diff --git a/server/core/lib/notifier/shared/caption/index.ts b/server/core/lib/notifier/shared/caption/index.ts
new file mode 100644
index 000000000..ec9db5564
--- /dev/null
+++ b/server/core/lib/notifier/shared/caption/index.ts
@@ -0,0 +1 @@
+export * from './video-transcription-generated-for-owner.js'
diff --git a/server/core/lib/notifier/shared/caption/video-transcription-generated-for-owner.ts b/server/core/lib/notifier/shared/caption/video-transcription-generated-for-owner.ts
new file mode 100644
index 000000000..a77740736
--- /dev/null
+++ b/server/core/lib/notifier/shared/caption/video-transcription-generated-for-owner.ts
@@ -0,0 +1,63 @@
+import { UserNotificationType } from '@peertube/peertube-models'
+import { logger } from '@server/helpers/logger.js'
+import { VIDEO_LANGUAGES, WEBSERVER } from '@server/initializers/constants.js'
+import { UserNotificationModel } from '@server/models/user/user-notification.js'
+import { UserModel } from '@server/models/user/user.js'
+import { MUserDefault, MUserWithNotificationSetting, MVideoCaptionVideo, UserNotificationModelForApi } from '@server/types/models/index.js'
+import { AbstractNotification } from '../common/abstract-notification.js'
+
+export class VideoTranscriptionGeneratedForOwner extends AbstractNotification <MVideoCaptionVideo> {
+  private user: MUserDefault
+
+  async prepare () {
+    this.user = await UserModel.loadByVideoId(this.payload.videoId)
+  }
+
+  log () {
+    logger.info(
+      'Notifying user %s the transcription %s of video %s is generated.',
+      this.user.username, this.payload.language, this.payload.Video.url
+    )
+  }
+
+  getSetting (user: MUserWithNotificationSetting) {
+    return user.NotificationSetting.myVideoTranscriptionGenerated
+  }
+
+  getTargetUsers () {
+    if (!this.user) return []
+
+    return [ this.user ]
+  }
+
+  createNotification (user: MUserWithNotificationSetting) {
+    const notification = UserNotificationModel.build<UserNotificationModelForApi>({
+      type: UserNotificationType.MY_VIDEO_TRANSCRIPTION_GENERATED,
+      userId: user.id,
+      videoCaptionId: this.payload.id
+    })
+    notification.VideoCaption = this.payload
+
+    return notification
+  }
+
+  createEmail (to: string) {
+    const video = this.payload.Video
+    const videoUrl = WEBSERVER.URL + video.getWatchStaticPath()
+
+    const language = VIDEO_LANGUAGES[this.payload.language]
+
+    return {
+      to,
+      subject: `Transcription in ${language} of your video ${video.name} has been generated`,
+      text: `Transcription in ${language} of your video ${video.name} has been generated.`,
+      locals: {
+        title: 'Transcription has been generated',
+        action: {
+          text: 'View video',
+          url: videoUrl
+        }
+      }
+    }
+  }
+}
diff --git a/server/core/lib/notifier/shared/index.ts b/server/core/lib/notifier/shared/index.ts
index f4f165c40..5ae060db7 100644
--- a/server/core/lib/notifier/shared/index.ts
+++ b/server/core/lib/notifier/shared/index.ts
@@ -1,5 +1,6 @@
 export * from './abuse/index.js'
 export * from './blacklist/index.js'
+export * from './caption/index.js'
 export * from './comment/index.js'
 export * from './common/index.js'
 export * from './follow/index.js'
diff --git a/server/core/lib/runners/job-handlers/abstract-job-handler.ts b/server/core/lib/runners/job-handlers/abstract-job-handler.ts
index d31effee6..55f7cd74a 100644
--- a/server/core/lib/runners/job-handlers/abstract-job-handler.ts
+++ b/server/core/lib/runners/job-handlers/abstract-job-handler.ts
@@ -6,6 +6,8 @@ import {
   RunnerJobStateType,
   RunnerJobStudioTranscodingPayload,
   RunnerJobSuccessPayload,
+  RunnerJobTranscriptionPayload,
+  RunnerJobTranscriptionPrivatePayload,
   RunnerJobType,
   RunnerJobUpdatePayload,
   RunnerJobVODAudioMergeTranscodingPayload,
@@ -51,6 +53,11 @@ type CreateRunnerJobArg =
     type: Extract<RunnerJobType, 'video-studio-transcoding'>
     payload: RunnerJobStudioTranscodingPayload
     privatePayload: RunnerJobVideoStudioTranscodingPrivatePayload
+  } |
+  {
+    type: Extract<RunnerJobType, 'video-transcription'>
+    payload: RunnerJobTranscriptionPayload
+    privatePayload: RunnerJobTranscriptionPrivatePayload
   }
 
 export abstract class AbstractJobHandler <C, U extends RunnerJobUpdatePayload, S extends RunnerJobSuccessPayload> {
diff --git a/server/core/lib/runners/job-handlers/abstract-vod-transcoding-job-handler.ts b/server/core/lib/runners/job-handlers/abstract-vod-transcoding-job-handler.ts
index d05a66ccf..20c0a62bb 100644
--- a/server/core/lib/runners/job-handlers/abstract-vod-transcoding-job-handler.ts
+++ b/server/core/lib/runners/job-handlers/abstract-vod-transcoding-job-handler.ts
@@ -11,7 +11,7 @@ import { moveToFailedTranscodingState, moveToNextState } from '@server/lib/video
 import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
 import { MRunnerJob } from '@server/types/models/runners/index.js'
 import { AbstractJobHandler } from './abstract-job-handler.js'
-import { loadTranscodingRunnerVideo } from './shared/index.js'
+import { loadRunnerVideo } from './shared/utils.js'
 
 // eslint-disable-next-line max-len
 export abstract class AbstractVODTranscodingJobHandler <C, U extends RunnerJobUpdatePayload, S extends RunnerJobSuccessPayload> extends AbstractJobHandler<C, U, S> {
@@ -38,7 +38,7 @@ export abstract class AbstractVODTranscodingJobHandler <C, U extends RunnerJobUp
   }) {
     if (options.nextState !== RunnerJobState.ERRORED) return
 
-    const video = await loadTranscodingRunnerVideo(options.runnerJob, this.lTags)
+    const video = await loadRunnerVideo(options.runnerJob, this.lTags)
     if (!video) return
 
     await moveToFailedTranscodingState(video)
@@ -51,7 +51,7 @@ export abstract class AbstractVODTranscodingJobHandler <C, U extends RunnerJobUp
   }) {
     const { runnerJob } = options
 
-    const video = await loadTranscodingRunnerVideo(options.runnerJob, this.lTags)
+    const video = await loadRunnerVideo(options.runnerJob, this.lTags)
     if (!video) return
 
     const pending = await VideoJobInfoModel.decrease(video.uuid, 'pendingTranscode')
diff --git a/server/core/lib/runners/job-handlers/index.ts b/server/core/lib/runners/job-handlers/index.ts
index c3205fae3..a9fb79160 100644
--- a/server/core/lib/runners/job-handlers/index.ts
+++ b/server/core/lib/runners/job-handlers/index.ts
@@ -1,6 +1,7 @@
 export * from './abstract-job-handler.js'
 export * from './live-rtmp-hls-transcoding-job-handler.js'
 export * from './runner-job-handlers.js'
+export * from './transcription-job-handler.js'
 export * from './video-studio-transcoding-job-handler.js'
 export * from './vod-audio-merge-transcoding-job-handler.js'
 export * from './vod-hls-transcoding-job-handler.js'
diff --git a/server/core/lib/runners/job-handlers/runner-job-handlers.ts b/server/core/lib/runners/job-handlers/runner-job-handlers.ts
index f8a672e64..771bf9d6a 100644
--- a/server/core/lib/runners/job-handlers/runner-job-handlers.ts
+++ b/server/core/lib/runners/job-handlers/runner-job-handlers.ts
@@ -1,7 +1,8 @@
-import { MRunnerJob } from '@server/types/models/runners/index.js'
 import { RunnerJobSuccessPayload, RunnerJobType, RunnerJobUpdatePayload } from '@peertube/peertube-models'
+import { MRunnerJob } from '@server/types/models/runners/index.js'
 import { AbstractJobHandler } from './abstract-job-handler.js'
 import { LiveRTMPHLSTranscodingJobHandler } from './live-rtmp-hls-transcoding-job-handler.js'
+import { TranscriptionJobHandler } from './transcription-job-handler.js'
 import { VideoStudioTranscodingJobHandler } from './video-studio-transcoding-job-handler.js'
 import { VODAudioMergeTranscodingJobHandler } from './vod-audio-merge-transcoding-job-handler.js'
 import { VODHLSTranscodingJobHandler } from './vod-hls-transcoding-job-handler.js'
@@ -12,7 +13,8 @@ const processors: Record<RunnerJobType, new() => AbstractJobHandler<unknown, Run
   'vod-hls-transcoding': VODHLSTranscodingJobHandler,
   'vod-audio-merge-transcoding': VODAudioMergeTranscodingJobHandler,
   'live-rtmp-hls-transcoding': LiveRTMPHLSTranscodingJobHandler,
-  'video-studio-transcoding': VideoStudioTranscodingJobHandler
+  'video-studio-transcoding': VideoStudioTranscodingJobHandler,
+  'video-transcription': TranscriptionJobHandler
 }
 
 export function getRunnerJobHandlerClass (job: MRunnerJob) {
diff --git a/server/core/lib/runners/job-handlers/shared/index.ts b/server/core/lib/runners/job-handlers/shared/index.ts
deleted file mode 100644
index bec61e3a0..000000000
--- a/server/core/lib/runners/job-handlers/shared/index.ts
+++ /dev/null
@@ -1 +0,0 @@
-export * from './vod-helpers.js'
diff --git a/server/core/lib/runners/job-handlers/shared/vod-helpers.ts b/server/core/lib/runners/job-handlers/shared/utils.ts
similarity index 87%
rename from server/core/lib/runners/job-handlers/shared/vod-helpers.ts
rename to server/core/lib/runners/job-handlers/shared/utils.ts
index 3c63206bc..e78d05235 100644
--- a/server/core/lib/runners/job-handlers/shared/vod-helpers.ts
+++ b/server/core/lib/runners/job-handlers/shared/utils.ts
@@ -23,12 +23,12 @@ export async function onVODWebVideoOrAudioMergeTranscodingJob (options: {
   await onTranscodingEnded({ isNewVideo: privatePayload.isNewVideo, moveVideoToNextState: true, video })
 }
 
-export async function loadTranscodingRunnerVideo (runnerJob: MRunnerJob, lTags: LoggerTagsFn) {
+export async function loadRunnerVideo (runnerJob: MRunnerJob, lTags: LoggerTagsFn) {
   const videoUUID = runnerJob.privatePayload.videoUUID
 
   const video = await VideoModel.loadFull(videoUUID)
   if (!video) {
-    logger.info('Video %s does not exist anymore after transcoding runner job.', videoUUID, lTags(videoUUID))
+    logger.info('Video %s does not exist anymore after runner job.', videoUUID, lTags(videoUUID))
     return undefined
   }
 
diff --git a/server/core/lib/runners/job-handlers/transcription-job-handler.ts b/server/core/lib/runners/job-handlers/transcription-job-handler.ts
new file mode 100644
index 000000000..077f1afc2
--- /dev/null
+++ b/server/core/lib/runners/job-handlers/transcription-job-handler.ts
@@ -0,0 +1,101 @@
+import {
+  RunnerJobState,
+  RunnerJobStateType,
+  RunnerJobTranscriptionPayload,
+  RunnerJobTranscriptionPrivatePayload,
+  RunnerJobUpdatePayload,
+  TranscriptionSuccess
+} from '@peertube/peertube-models'
+import { buildUUID } from '@peertube/peertube-node-utils'
+import { JOB_PRIORITY } from '@server/initializers/constants.js'
+import { onTranscriptionEnded } from '@server/lib/video-captions.js'
+import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
+import { MVideoUUID } from '@server/types/models/index.js'
+import { MRunnerJob } from '@server/types/models/runners/index.js'
+import { generateRunnerTranscodingVideoInputFileUrl } from '../runner-urls.js'
+import { AbstractJobHandler } from './abstract-job-handler.js'
+import { loadRunnerVideo } from './shared/utils.js'
+
+type CreateOptions = {
+  video: MVideoUUID
+}
+
+export class TranscriptionJobHandler extends AbstractJobHandler<CreateOptions, RunnerJobUpdatePayload, TranscriptionSuccess> {
+
+  protected isAbortSupported () {
+    return true
+  }
+
+  protected specificUpdate (_options: {
+    runnerJob: MRunnerJob
+  }) {
+    // empty
+  }
+
+  protected specificAbort (_options: {
+    runnerJob: MRunnerJob
+  }) {
+    // empty
+  }
+
+  protected async specificError (options: {
+    runnerJob: MRunnerJob
+    nextState: RunnerJobStateType
+  }) {
+    if (options.nextState !== RunnerJobState.ERRORED) return
+
+    await VideoJobInfoModel.decrease(options.runnerJob.privatePayload.videoUUID, 'pendingTranscription')
+  }
+
+  protected async specificCancel (options: {
+    runnerJob: MRunnerJob
+  }) {
+    await VideoJobInfoModel.decrease(options.runnerJob.privatePayload.videoUUID, 'pendingTranscription')
+  }
+
+  async create (options: CreateOptions) {
+    const { video } = options
+
+    const jobUUID = buildUUID()
+    const payload: RunnerJobTranscriptionPayload = {
+      input: {
+        videoFileUrl: generateRunnerTranscodingVideoInputFileUrl(jobUUID, video.uuid)
+      }
+    }
+
+    const privatePayload: RunnerJobTranscriptionPrivatePayload = {
+      videoUUID: video.uuid
+    }
+
+    const job = await this.createRunnerJob({
+      type: 'video-transcription',
+      jobUUID,
+      payload,
+      privatePayload,
+      priority: JOB_PRIORITY.TRANSCODING
+    })
+
+    await VideoJobInfoModel.increaseOrCreate(video.uuid, 'pendingTranscription')
+
+    return job
+  }
+
+  // ---------------------------------------------------------------------------
+
+  protected async specificComplete (options: {
+    runnerJob: MRunnerJob
+    resultPayload: TranscriptionSuccess
+  }) {
+    const { runnerJob, resultPayload } = options
+
+    const video = await loadRunnerVideo(runnerJob, this.lTags)
+    if (!video) return
+
+    await onTranscriptionEnded({
+      video,
+      language: resultPayload.inputLanguage,
+      vttPath: resultPayload.vttFile as string,
+      lTags: this.lTags().tags
+    })
+  }
+}
diff --git a/server/core/lib/runners/job-handlers/video-studio-transcoding-job-handler.ts b/server/core/lib/runners/job-handlers/video-studio-transcoding-job-handler.ts
index d958d447f..a244cddff 100644
--- a/server/core/lib/runners/job-handlers/video-studio-transcoding-job-handler.ts
+++ b/server/core/lib/runners/job-handlers/video-studio-transcoding-job-handler.ts
@@ -19,7 +19,7 @@ import { MRunnerJob } from '@server/types/models/runners/index.js'
 import { basename } from 'path'
 import { generateRunnerEditionTranscodingVideoInputFileUrl, generateRunnerTranscodingVideoInputFileUrl } from '../runner-urls.js'
 import { AbstractJobHandler } from './abstract-job-handler.js'
-import { loadTranscodingRunnerVideo } from './shared/index.js'
+import { loadRunnerVideo } from './shared/utils.js'
 
 type CreateOptions = {
   video: MVideo
@@ -108,7 +108,7 @@ export class VideoStudioTranscodingJobHandler extends AbstractJobHandler<CreateO
     const { runnerJob, resultPayload } = options
     const privatePayload = runnerJob.privatePayload as RunnerJobVideoStudioTranscodingPrivatePayload
 
-    const video = await loadTranscodingRunnerVideo(runnerJob, this.lTags)
+    const video = await loadRunnerVideo(runnerJob, this.lTags)
     if (!video) {
       await safeCleanupStudioTMPFiles(privatePayload.originalTasks)
 
@@ -149,7 +149,7 @@ export class VideoStudioTranscodingJobHandler extends AbstractJobHandler<CreateO
     const payload = runnerJob.privatePayload as RunnerJobVideoStudioTranscodingPrivatePayload
     await safeCleanupStudioTMPFiles(payload.originalTasks)
 
-    const video = await loadTranscodingRunnerVideo(options.runnerJob, this.lTags)
+    const video = await loadRunnerVideo(options.runnerJob, this.lTags)
     if (!video) return
 
     return video.setNewState(VideoState.PUBLISHED, false, undefined)
diff --git a/server/core/lib/runners/job-handlers/vod-audio-merge-transcoding-job-handler.ts b/server/core/lib/runners/job-handlers/vod-audio-merge-transcoding-job-handler.ts
index b93ab37a4..5933501f5 100644
--- a/server/core/lib/runners/job-handlers/vod-audio-merge-transcoding-job-handler.ts
+++ b/server/core/lib/runners/job-handlers/vod-audio-merge-transcoding-job-handler.ts
@@ -1,18 +1,18 @@
-import { logger } from '@server/helpers/logger.js'
-import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
-import { MVideo } from '@server/types/models/index.js'
-import { MRunnerJob } from '@server/types/models/runners/index.js'
 import { pick } from '@peertube/peertube-core-utils'
-import { buildUUID } from '@peertube/peertube-node-utils'
 import {
   RunnerJobUpdatePayload,
   RunnerJobVODAudioMergeTranscodingPayload,
   RunnerJobVODWebVideoTranscodingPrivatePayload,
   VODAudioMergeTranscodingSuccess
 } from '@peertube/peertube-models'
+import { buildUUID } from '@peertube/peertube-node-utils'
+import { logger } from '@server/helpers/logger.js'
+import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
+import { MVideo } from '@server/types/models/index.js'
+import { MRunnerJob } from '@server/types/models/runners/index.js'
 import { generateRunnerTranscodingVideoInputFileUrl, generateRunnerTranscodingVideoPreviewFileUrl } from '../runner-urls.js'
 import { AbstractVODTranscodingJobHandler } from './abstract-vod-transcoding-job-handler.js'
-import { loadTranscodingRunnerVideo, onVODWebVideoOrAudioMergeTranscodingJob } from './shared/index.js'
+import { loadRunnerVideo, onVODWebVideoOrAudioMergeTranscodingJob } from './shared/utils.js'
 
 type CreateOptions = {
   video: MVideo
@@ -71,7 +71,7 @@ export class VODAudioMergeTranscodingJobHandler extends AbstractVODTranscodingJo
     const { runnerJob, resultPayload } = options
     const privatePayload = runnerJob.privatePayload as RunnerJobVODWebVideoTranscodingPrivatePayload
 
-    const video = await loadTranscodingRunnerVideo(runnerJob, this.lTags)
+    const video = await loadRunnerVideo(runnerJob, this.lTags)
     if (!video) return
 
     const videoFilePath = resultPayload.videoFile as string
diff --git a/server/core/lib/runners/job-handlers/vod-hls-transcoding-job-handler.ts b/server/core/lib/runners/job-handlers/vod-hls-transcoding-job-handler.ts
index 0cb011d95..283ca7c30 100644
--- a/server/core/lib/runners/job-handlers/vod-hls-transcoding-job-handler.ts
+++ b/server/core/lib/runners/job-handlers/vod-hls-transcoding-job-handler.ts
@@ -1,3 +1,11 @@
+import { pick } from '@peertube/peertube-core-utils'
+import {
+  RunnerJobUpdatePayload,
+  RunnerJobVODHLSTranscodingPayload,
+  RunnerJobVODHLSTranscodingPrivatePayload,
+  VODHLSTranscodingSuccess
+} from '@peertube/peertube-models'
+import { buildUUID } from '@peertube/peertube-node-utils'
 import { logger } from '@server/helpers/logger.js'
 import { onTranscodingEnded } from '@server/lib/transcoding/ended-transcoding.js'
 import { onHLSVideoFileTranscoding } from '@server/lib/transcoding/hls-transcoding.js'
@@ -5,17 +13,9 @@ import { removeAllWebVideoFiles } from '@server/lib/video-file.js'
 import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
 import { MVideo } from '@server/types/models/index.js'
 import { MRunnerJob } from '@server/types/models/runners/index.js'
-import { pick } from '@peertube/peertube-core-utils'
-import { buildUUID } from '@peertube/peertube-node-utils'
-import {
-  RunnerJobUpdatePayload,
-  RunnerJobVODHLSTranscodingPayload,
-  RunnerJobVODHLSTranscodingPrivatePayload,
-  VODHLSTranscodingSuccess
-} from '@peertube/peertube-models'
 import { generateRunnerTranscodingVideoInputFileUrl } from '../runner-urls.js'
 import { AbstractVODTranscodingJobHandler } from './abstract-vod-transcoding-job-handler.js'
-import { loadTranscodingRunnerVideo } from './shared/index.js'
+import { loadRunnerVideo } from './shared/utils.js'
 
 type CreateOptions = {
   video: MVideo
@@ -74,7 +74,7 @@ export class VODHLSTranscodingJobHandler extends AbstractVODTranscodingJobHandle
     const { runnerJob, resultPayload } = options
     const privatePayload = runnerJob.privatePayload as RunnerJobVODHLSTranscodingPrivatePayload
 
-    const video = await loadTranscodingRunnerVideo(runnerJob, this.lTags)
+    const video = await loadRunnerVideo(runnerJob, this.lTags)
     if (!video) return
 
     const videoFilePath = resultPayload.videoFile as string
diff --git a/server/core/lib/runners/job-handlers/vod-web-video-transcoding-job-handler.ts b/server/core/lib/runners/job-handlers/vod-web-video-transcoding-job-handler.ts
index 12a846985..081995a48 100644
--- a/server/core/lib/runners/job-handlers/vod-web-video-transcoding-job-handler.ts
+++ b/server/core/lib/runners/job-handlers/vod-web-video-transcoding-job-handler.ts
@@ -1,18 +1,18 @@
-import { logger } from '@server/helpers/logger.js'
-import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
-import { MVideo } from '@server/types/models/index.js'
-import { MRunnerJob } from '@server/types/models/runners/index.js'
 import { pick } from '@peertube/peertube-core-utils'
-import { buildUUID } from '@peertube/peertube-node-utils'
 import {
   RunnerJobUpdatePayload,
   RunnerJobVODWebVideoTranscodingPayload,
   RunnerJobVODWebVideoTranscodingPrivatePayload,
   VODWebVideoTranscodingSuccess
 } from '@peertube/peertube-models'
+import { buildUUID } from '@peertube/peertube-node-utils'
+import { logger } from '@server/helpers/logger.js'
+import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
+import { MVideo } from '@server/types/models/index.js'
+import { MRunnerJob } from '@server/types/models/runners/index.js'
 import { generateRunnerTranscodingVideoInputFileUrl } from '../runner-urls.js'
 import { AbstractVODTranscodingJobHandler } from './abstract-vod-transcoding-job-handler.js'
-import { loadTranscodingRunnerVideo, onVODWebVideoOrAudioMergeTranscodingJob } from './shared/index.js'
+import { loadRunnerVideo, onVODWebVideoOrAudioMergeTranscodingJob } from './shared/utils.js'
 
 type CreateOptions = {
   video: MVideo
@@ -70,7 +70,7 @@ export class VODWebVideoTranscodingJobHandler extends AbstractVODTranscodingJobH
     const { runnerJob, resultPayload } = options
     const privatePayload = runnerJob.privatePayload as RunnerJobVODWebVideoTranscodingPrivatePayload
 
-    const video = await loadTranscodingRunnerVideo(runnerJob, this.lTags)
+    const video = await loadRunnerVideo(runnerJob, this.lTags)
     if (!video) return
 
     const videoFilePath = resultPayload.videoFile as string
diff --git a/server/core/lib/server-config-manager.ts b/server/core/lib/server-config-manager.ts
index 523c2e554..72e4ac24f 100644
--- a/server/core/lib/server-config-manager.ts
+++ b/server/core/lib/server-config-manager.ts
@@ -192,6 +192,9 @@ class ServerConfigManager {
           enabled: CONFIG.VIDEO_FILE.UPDATE.ENABLED
         }
       },
+      videoTranscription: {
+        enabled: CONFIG.VIDEO_TRANSCRIPTION.ENABLED
+      },
       import: {
         videos: {
           http: {
diff --git a/server/core/lib/user-import-export/importers/user-settings-importer.ts b/server/core/lib/user-import-export/importers/user-settings-importer.ts
index 738f81237..fd72a46ee 100644
--- a/server/core/lib/user-import-export/importers/user-settings-importer.ts
+++ b/server/core/lib/user-import-export/importers/user-settings-importer.ts
@@ -93,7 +93,8 @@ export class UserSettingsImporter extends AbstractUserImporter <UserSettingsExpo
       autoInstanceFollowing: settingsImportData.autoInstanceFollowing,
       newPeerTubeVersion: settingsImportData.newPeerTubeVersion,
       newPluginVersion: settingsImportData.newPluginVersion,
-      myVideoStudioEditionFinished: settingsImportData.myVideoStudioEditionFinished
+      myVideoStudioEditionFinished: settingsImportData.myVideoStudioEditionFinished,
+      myVideoTranscriptionGenerated: settingsImportData.myVideoTranscriptionGenerated
     }
 
     await UserNotificationSettingModel.updateUserSettings(values, this.user.id)
diff --git a/server/core/lib/user.ts b/server/core/lib/user.ts
index 3c3b0de0f..39ec281a6 100644
--- a/server/core/lib/user.ts
+++ b/server/core/lib/user.ts
@@ -274,7 +274,8 @@ function createDefaultUserNotificationSettings (user: MUserId, t: Transaction |
     autoInstanceFollowing: UserNotificationSettingValue.WEB,
     newPeerTubeVersion: UserNotificationSettingValue.WEB | UserNotificationSettingValue.EMAIL,
     newPluginVersion: UserNotificationSettingValue.WEB,
-    myVideoStudioEditionFinished: UserNotificationSettingValue.WEB
+    myVideoStudioEditionFinished: UserNotificationSettingValue.WEB,
+    myVideoTranscriptionGenerated: UserNotificationSettingValue.WEB
   }
 
   return UserNotificationSettingModel.create(values, { transaction: t })
diff --git a/server/core/lib/video-captions.ts b/server/core/lib/video-captions.ts
index 8b363274f..ddd1418dc 100644
--- a/server/core/lib/video-captions.ts
+++ b/server/core/lib/video-captions.ts
@@ -1,7 +1,25 @@
+import { hasAudioStream } from '@peertube/peertube-ffmpeg'
+import { buildSUUID } from '@peertube/peertube-node-utils'
+import { AbstractTranscriber, TranscriptionModel, WhisperBuiltinModel, transcriberFactory } from '@peertube/peertube-transcription'
 import { moveAndProcessCaptionFile } from '@server/helpers/captions-utils.js'
+import { isVideoCaptionLanguageValid } from '@server/helpers/custom-validators/video-captions.js'
+import { logger, loggerTagsFactory } from '@server/helpers/logger.js'
+import { CONFIG } from '@server/initializers/config.js'
+import { DIRECTORIES } from '@server/initializers/constants.js'
 import { sequelizeTypescript } from '@server/initializers/database.js'
 import { VideoCaptionModel } from '@server/models/video/video-caption.js'
-import { MVideo, MVideoCaption } from '@server/types/models/index.js'
+import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
+import { VideoModel } from '@server/models/video/video.js'
+import { MVideo, MVideoCaption, MVideoFullLight, MVideoUUID, MVideoUrl } from '@server/types/models/index.js'
+import { ensureDir, remove } from 'fs-extra/esm'
+import { join } from 'path'
+import { federateVideoIfNeeded } from './activitypub/videos/federate.js'
+import { JobQueue } from './job-queue/job-queue.js'
+import { Notifier } from './notifier/notifier.js'
+import { TranscriptionJobHandler } from './runners/index.js'
+import { VideoPathManager } from './video-path-manager.js'
+
+const lTags = loggerTagsFactory('video-caption')
 
 export async function createLocalCaption (options: {
   video: MVideo
@@ -22,5 +40,122 @@ export async function createLocalCaption (options: {
     await VideoCaptionModel.insertOrReplaceLanguage(videoCaption, t)
   })
 
-  return videoCaption
+  return Object.assign(videoCaption, { Video: video })
+}
+
+export async function createTranscriptionTaskIfNeeded (video: MVideoUUID & MVideoUrl) {
+  if (CONFIG.VIDEO_TRANSCRIPTION.ENABLED !== true) return
+
+  logger.info(`Creating transcription job for ${video.url}`, lTags(video.uuid))
+
+  if (CONFIG.VIDEO_TRANSCRIPTION.REMOTE_RUNNERS.ENABLED === true) {
+    await new TranscriptionJobHandler().create({ video })
+  } else {
+    await JobQueue.Instance.createJob({ type: 'video-transcription', payload: { videoUUID: video.uuid } })
+  }
+
+  await VideoJobInfoModel.increaseOrCreate(video.uuid, 'pendingTranscription')
+}
+
+// ---------------------------------------------------------------------------
+// Transcription task
+// ---------------------------------------------------------------------------
+
+let transcriber: AbstractTranscriber
+
+export async function generateSubtitle (options: {
+  video: MVideoUUID
+}) {
+  const inputFileMutexReleaser = await VideoPathManager.Instance.lockFiles(options.video.uuid)
+
+  const outputPath = join(CONFIG.STORAGE.TMP_DIR, 'transcription', buildSUUID())
+  await ensureDir(outputPath)
+
+  const binDirectory = join(DIRECTORIES.LOCAL_PIP_DIRECTORY, 'bin')
+
+  try {
+    // Lazy load the transcriber
+    if (!transcriber) {
+      transcriber = transcriberFactory.createFromEngineName({
+        engineName: CONFIG.VIDEO_TRANSCRIPTION.ENGINE,
+        enginePath: CONFIG.VIDEO_TRANSCRIPTION.ENGINE_PATH,
+        logger,
+        binDirectory
+      })
+
+      if (!CONFIG.VIDEO_TRANSCRIPTION.ENGINE_PATH) {
+        logger.info(`Installing transcriber ${transcriber.engine.name} to generate subtitles`, lTags())
+        await transcriber.install(DIRECTORIES.LOCAL_PIP_DIRECTORY)
+      }
+    }
+
+    const video = await VideoModel.loadFull(options.video.uuid)
+    const file = video.getMaxQualityFile().withVideoOrPlaylist(video)
+
+    await VideoPathManager.Instance.makeAvailableVideoFile(file, async videoInputPath => {
+      if (await hasAudioStream(videoInputPath) !== true) {
+        logger.info(
+          `Do not run transcription for ${video.uuid} in ${outputPath} because it does not contain an audio stream`,
+          lTags(video.uuid)
+        )
+
+        return
+      }
+
+      logger.info(`Running transcription for ${video.uuid} in ${outputPath}`, lTags(video.uuid))
+
+      const transcriptFile = await transcriber.transcribe({
+        mediaFilePath: videoInputPath,
+
+        model: CONFIG.VIDEO_TRANSCRIPTION.MODEL_PATH
+          ? await TranscriptionModel.fromPath(CONFIG.VIDEO_TRANSCRIPTION.MODEL_PATH)
+          : new WhisperBuiltinModel(CONFIG.VIDEO_TRANSCRIPTION.MODEL),
+
+        transcriptDirectory: outputPath,
+
+        format: 'vtt'
+      })
+
+      await onTranscriptionEnded({ video, language: transcriptFile.language, vttPath: transcriptFile.path })
+    })
+  } finally {
+    if (outputPath) await remove(outputPath)
+
+    inputFileMutexReleaser()
+  }
+}
+
+export async function onTranscriptionEnded (options: {
+  video: MVideoFullLight
+  language: string
+  vttPath: string
+  lTags?: (string | number)[]
+}) {
+  const { video, language, vttPath, lTags: customLTags = [] } = options
+
+  await VideoJobInfoModel.decrease(video.uuid, 'pendingTranscription')
+
+  if (!isVideoCaptionLanguageValid(language)) {
+    logger.warn(`Invalid transcription language for video ${video.uuid}`, this.lTags(video.uuid))
+    return
+  }
+
+  if (!video.language) {
+    video.language = language
+    await video.save()
+  }
+
+  const caption = await createLocalCaption({
+    video,
+    language,
+    path: vttPath
+  })
+
+  await sequelizeTypescript.transaction(async t => {
+    await federateVideoIfNeeded(video, false, t)
+  })
+
+  Notifier.Instance.notifyOfGeneratedVideoTranscription(caption)
+
+  logger.info(`Transcription ended for ${video.uuid}`, lTags(video.uuid, ...customLTags))
 }
diff --git a/server/core/lib/video-file.ts b/server/core/lib/video-file.ts
index 5134f76c9..b43176fb1 100644
--- a/server/core/lib/video-file.ts
+++ b/server/core/lib/video-file.ts
@@ -8,7 +8,7 @@ import { VideoFileModel } from '@server/models/video/video-file.js'
 import { VideoSourceModel } from '@server/models/video/video-source.js'
 import { MVideo, MVideoFile, MVideoId, MVideoWithAllFiles } from '@server/types/models/index.js'
 import { FfprobeData } from 'fluent-ffmpeg'
-import { move, remove } from 'fs-extra'
+import { move, remove } from 'fs-extra/esm'
 import { lTags } from './object-storage/shared/index.js'
 import { storeOriginalVideoFile } from './object-storage/videos.js'
 import { generateHLSVideoFilename, generateWebVideoFilename } from './paths.js'
diff --git a/server/core/lib/video-jobs.ts b/server/core/lib/video-jobs.ts
index d69cda8a3..9ad05a4c4 100644
--- a/server/core/lib/video-jobs.ts
+++ b/server/core/lib/video-jobs.ts
@@ -3,6 +3,7 @@ import { CONFIG } from '@server/initializers/config.js'
 import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
 import { MVideo, MVideoFile, MVideoFullLight, MVideoUUID } from '@server/types/models/index.js'
 import { CreateJobArgument, CreateJobOptions, JobQueue } from './job-queue/job-queue.js'
+import { createTranscriptionTaskIfNeeded } from './video-captions.js'
 import { moveFilesIfPrivacyChanged } from './video-privacy.js'
 
 export async function buildMoveJob (options: {
@@ -57,8 +58,9 @@ export function buildStoryboardJobIfNeeded (options: {
 export async function addVideoJobsAfterCreation (options: {
   video: MVideo
   videoFile: MVideoFile
+  generateTranscription: boolean
 }) {
-  const { video, videoFile } = options
+  const { video, videoFile, generateTranscription } = options
 
   const jobs: (CreateJobArgument & CreateJobOptions)[] = [
     {
@@ -105,7 +107,11 @@ export async function addVideoJobsAfterCreation (options: {
     })
   }
 
-  return JobQueue.Instance.createSequentialJobFlow(...jobs)
+  await JobQueue.Instance.createSequentialJobFlow(...jobs)
+
+  if (generateTranscription === true && CONFIG.VIDEO_TRANSCRIPTION.ENABLED === true) {
+    await createTranscriptionTaskIfNeeded(video)
+  }
 }
 
 export async function addVideoJobsAfterUpdate (options: {
diff --git a/server/core/lib/video-pre-import.ts b/server/core/lib/video-pre-import.ts
index ceca6580b..fbb94edac 100644
--- a/server/core/lib/video-pre-import.ts
+++ b/server/core/lib/video-pre-import.ts
@@ -259,6 +259,7 @@ async function buildYoutubeDLImport (options: {
     type: 'youtube-dl' as 'youtube-dl',
     videoImportId: videoImport.id,
     fileExt,
+    generateTranscription: importDataOverride.generateTranscription ?? true,
     // If part of a sync process, there is a parent job that will aggregate children results
     preventException: !!channelSync
   }
diff --git a/server/core/middlewares/validators/videos/video-captions.ts b/server/core/middlewares/validators/videos/video-captions.ts
index d6407502b..74ad1e8aa 100644
--- a/server/core/middlewares/validators/videos/video-captions.ts
+++ b/server/core/middlewares/validators/videos/video-captions.ts
@@ -1,6 +1,10 @@
+import { HttpStatusCode, ServerErrorCode, UserRight, VideoCaptionGenerate } from '@peertube/peertube-models'
+import { isBooleanValid, toBooleanOrNull } from '@server/helpers/custom-validators/misc.js'
+import { CONFIG } from '@server/initializers/config.js'
+import { VideoCaptionModel } from '@server/models/video/video-caption.js'
+import { VideoJobInfoModel } from '@server/models/video/video-job-info.js'
 import express from 'express'
 import { body, param } from 'express-validator'
-import { UserRight } from '@peertube/peertube-models'
 import { isVideoCaptionFile, isVideoCaptionLanguageValid } from '../../../helpers/custom-validators/video-captions.js'
 import { cleanUpReqFiles } from '../../../helpers/express-utils.js'
 import { CONSTRAINTS_FIELDS, MIMETYPES } from '../../../initializers/constants.js'
@@ -14,7 +18,7 @@ import {
   isValidVideoPasswordHeader
 } from '../shared/index.js'
 
-const addVideoCaptionValidator = [
+export const addVideoCaptionValidator = [
   isValidVideoIdParam('videoId'),
 
   param('captionLanguage')
@@ -41,7 +45,75 @@ const addVideoCaptionValidator = [
   }
 ]
 
-const deleteVideoCaptionValidator = [
+export const generateVideoCaptionValidator = [
+  isValidVideoIdParam('videoId'),
+
+  body('forceTranscription')
+    .optional()
+    .customSanitizer(toBooleanOrNull)
+    .custom(isBooleanValid),
+
+  async (req: express.Request, res: express.Response, next: express.NextFunction) => {
+    if (areValidationErrors(req, res)) return
+
+    if (CONFIG.VIDEO_TRANSCRIPTION.ENABLED !== true) {
+      return res.fail({
+        status: HttpStatusCode.BAD_REQUEST_400,
+        message: 'Video transcription is disabled on this instance'
+      })
+    }
+
+    if (!await doesVideoExist(req.params.videoId, res)) return
+
+    const video = res.locals.videoAll
+
+    if (video.remote) {
+      return res.fail({
+        status: HttpStatusCode.BAD_REQUEST_400,
+        message: 'Cannot run transcription job on a remote video'
+      })
+    }
+
+    // Check if the user who did the request is able to update the video
+    const user = res.locals.oauth.token.User
+    if (!checkUserCanManageVideo(user, video, UserRight.UPDATE_ANY_VIDEO, res)) return
+
+    // Check the video has not already a caption
+    const captions = await VideoCaptionModel.listVideoCaptions(video.id)
+    if (captions.length !== 0) {
+      return res.fail({
+        status: HttpStatusCode.BAD_REQUEST_400,
+        message: 'This video already has captions'
+      })
+    }
+
+    // Bypass "video is already transcribed" check
+    const body = req.body as VideoCaptionGenerate
+    if (body.forceTranscription === true) {
+      if (user.hasRight(UserRight.UPDATE_ANY_VIDEO) !== true) {
+        return res.fail({
+          status: HttpStatusCode.FORBIDDEN_403,
+          message: 'Only admins can force transcription'
+        })
+      }
+
+      return next()
+    }
+
+    const info = await VideoJobInfoModel.load(video.id)
+    if (info && info.pendingTranscription > 0) {
+      return res.fail({
+        status: HttpStatusCode.CONFLICT_409,
+        type: ServerErrorCode.VIDEO_ALREADY_BEING_TRANSCRIBED,
+        message: 'This video is already being transcribed'
+      })
+    }
+
+    return next()
+  }
+]
+
+export const deleteVideoCaptionValidator = [
   isValidVideoIdParam('videoId'),
 
   param('captionLanguage')
@@ -60,7 +132,7 @@ const deleteVideoCaptionValidator = [
   }
 ]
 
-const listVideoCaptionsValidator = [
+export const listVideoCaptionsValidator = [
   isValidVideoIdParam('videoId'),
 
   isValidVideoPasswordHeader(),
@@ -75,9 +147,3 @@ const listVideoCaptionsValidator = [
     return next()
   }
 ]
-
-export {
-  addVideoCaptionValidator,
-  listVideoCaptionsValidator,
-  deleteVideoCaptionValidator
-}
diff --git a/server/core/models/user/sql/user-notitication-list-query-builder.ts b/server/core/models/user/sql/user-notitication-list-query-builder.ts
index 32f5ba73e..fe0720e76 100644
--- a/server/core/models/user/sql/user-notitication-list-query-builder.ts
+++ b/server/core/models/user/sql/user-notitication-list-query-builder.ts
@@ -183,7 +183,12 @@ export class UserNotificationListQueryBuilder extends AbstractRunQuery {
       "Account->Actor->Server"."id" AS "Account.Actor.Server.id",
       "Account->Actor->Server"."host" AS "Account.Actor.Server.host",
       "UserRegistration"."id" AS "UserRegistration.id",
-      "UserRegistration"."username" AS "UserRegistration.username"`
+      "UserRegistration"."username" AS "UserRegistration.username",
+      "VideoCaption"."id" AS "VideoCaption.id",
+      "VideoCaption"."language" AS "VideoCaption.language",
+      "VideoCaption->Video"."id" AS "VideoCaption.Video.id",
+      "VideoCaption->Video"."uuid" AS "VideoCaption.Video.uuid",
+      "VideoCaption->Video"."name" AS "VideoCaption.Video.name"`
   }
 
   private getJoins () {
@@ -269,6 +274,11 @@ export class UserNotificationListQueryBuilder extends AbstractRunQuery {
       LEFT JOIN "server" AS "Account->Actor->Server" ON "Account->Actor"."serverId" = "Account->Actor->Server"."id"
     ) ON "UserNotificationModel"."accountId" = "Account"."id"
 
-    LEFT JOIN "userRegistration" as "UserRegistration" ON "UserNotificationModel"."userRegistrationId" = "UserRegistration"."id"`
+    LEFT JOIN "userRegistration" as "UserRegistration" ON "UserNotificationModel"."userRegistrationId" = "UserRegistration"."id"
+
+    LEFT JOIN (
+      "videoCaption" AS "VideoCaption"
+      INNER JOIN "video" AS "VideoCaption->Video" ON "VideoCaption"."videoId" = "VideoCaption->Video"."id"
+    ) ON "UserNotificationModel"."videoCaptionId" = "VideoCaption"."id"`
   }
 }
diff --git a/server/core/models/user/user-notification-setting.ts b/server/core/models/user/user-notification-setting.ts
index 2ad3ddfa9..dc48ca55d 100644
--- a/server/core/models/user/user-notification-setting.ts
+++ b/server/core/models/user/user-notification-setting.ts
@@ -181,6 +181,15 @@ export class UserNotificationSettingModel extends SequelizeModel<UserNotificatio
   @Column
   myVideoStudioEditionFinished: UserNotificationSettingValueType
 
+  @AllowNull(false)
+  @Default(null)
+  @Is(
+    'UserNotificationSettingTranscriptionGeneratedForOwner',
+    value => throwIfNotValid(value, isUserNotificationSettingValid, 'myVideoTranscriptionGenerated')
+  )
+  @Column
+  myVideoTranscriptionGenerated: UserNotificationSettingValueType
+
   @ForeignKey(() => UserModel)
   @Column
   userId: number
@@ -233,6 +242,7 @@ export class UserNotificationSettingModel extends SequelizeModel<UserNotificatio
       abuseStateChange: this.abuseStateChange,
       newPeerTubeVersion: this.newPeerTubeVersion,
       myVideoStudioEditionFinished: this.myVideoStudioEditionFinished,
+      myVideoTranscriptionGenerated: this.myVideoTranscriptionGenerated,
       newPluginVersion: this.newPluginVersion
     }
   }
diff --git a/server/core/models/user/user-notification.ts b/server/core/models/user/user-notification.ts
index 0b41bf676..d1cd9c123 100644
--- a/server/core/models/user/user-notification.ts
+++ b/server/core/models/user/user-notification.ts
@@ -13,6 +13,7 @@ import { ApplicationModel } from '../application/application.js'
 import { PluginModel } from '../server/plugin.js'
 import { SequelizeModel, throwIfNotValid } from '../shared/index.js'
 import { VideoBlacklistModel } from '../video/video-blacklist.js'
+import { VideoCaptionModel } from '../video/video-caption.js'
 import { VideoCommentModel } from '../video/video-comment.js'
 import { VideoImportModel } from '../video/video-import.js'
 import { VideoModel } from '../video/video.js'
@@ -260,6 +261,18 @@ export class UserNotificationModel extends SequelizeModel<UserNotificationModel>
   })
   UserRegistration: Awaited<UserRegistrationModel>
 
+  @ForeignKey(() => VideoCaptionModel)
+  @Column
+  videoCaptionId: number
+
+  @BelongsTo(() => VideoCaptionModel, {
+    foreignKey: {
+      allowNull: true
+    },
+    onDelete: 'cascade'
+  })
+  VideoCaption: Awaited<VideoCaptionModel>
+
   static listForApi (userId: number, start: number, count: number, sort: string, unread?: boolean) {
     const where = { userId }
 
@@ -441,6 +454,17 @@ export class UserNotificationModel extends SequelizeModel<UserNotificationModel>
       ? { id: this.UserRegistration.id, username: this.UserRegistration.username }
       : undefined
 
+    const videoCaption = this.VideoCaption
+      ? {
+        id: this.VideoCaption.id,
+        language: {
+          id: this.VideoCaption.language,
+          label: VideoCaptionModel.getLanguageLabel(this.VideoCaption.language)
+        },
+        video: this.formatVideo(this.VideoCaption.Video)
+      }
+      : undefined
+
     return {
       id: this.id,
       type: this.type,
@@ -455,6 +479,7 @@ export class UserNotificationModel extends SequelizeModel<UserNotificationModel>
       plugin,
       peertube,
       registration,
+      videoCaption,
       createdAt: this.createdAt.toISOString(),
       updatedAt: this.updatedAt.toISOString()
     }
@@ -531,4 +556,11 @@ export class UserNotificationModel extends SequelizeModel<UserNotificationModel>
       width: a.width
     }
   }
+
+  formatVideoCaption (a: UserNotificationIncludes.ActorImageInclude) {
+    return {
+      path: a.getStaticPath(),
+      width: a.width
+    }
+  }
 }
diff --git a/server/core/models/video/video-caption.ts b/server/core/models/video/video-caption.ts
index e707e868f..df2124b86 100644
--- a/server/core/models/video/video-caption.ts
+++ b/server/core/models/video/video-caption.ts
@@ -112,10 +112,21 @@ export class VideoCaptionModel extends SequelizeModel<VideoCaptionModel> {
     return undefined
   }
 
+  static async insertOrReplaceLanguage (caption: MVideoCaption, transaction: Transaction) {
+    const existing = await VideoCaptionModel.loadByVideoIdAndLanguage(caption.videoId, caption.language, transaction)
+
+    // Delete existing file
+    if (existing) await existing.destroy({ transaction })
+
+    return caption.save({ transaction })
+  }
+
+  // ---------------------------------------------------------------------------
+
   static loadByVideoIdAndLanguage (videoId: string | number, language: string, transaction?: Transaction): Promise<MVideoCaptionVideo> {
     const videoInclude = {
       model: VideoModel.unscoped(),
-      attributes: [ 'id', 'remote', 'uuid' ],
+      attributes: [ 'id', 'name', 'remote', 'uuid', 'url' ],
       where: buildWhereIdOrUUID(videoId)
     }
 
@@ -148,13 +159,18 @@ export class VideoCaptionModel extends SequelizeModel<VideoCaptionModel> {
     return VideoCaptionModel.findOne(query)
   }
 
-  static async insertOrReplaceLanguage (caption: MVideoCaption, transaction: Transaction) {
-    const existing = await VideoCaptionModel.loadByVideoIdAndLanguage(caption.videoId, caption.language, transaction)
+  // ---------------------------------------------------------------------------
 
-    // Delete existing file
-    if (existing) await existing.destroy({ transaction })
+  static async hasVideoCaption (videoId: number) {
+    const query = {
+      where: {
+        videoId
+      }
+    }
 
-    return caption.save({ transaction })
+    const result = await VideoCaptionModel.unscoped().findOne(query)
+
+    return !!result
   }
 
   static listVideoCaptions (videoId: number, transaction?: Transaction): Promise<MVideoCaptionVideo[]> {
@@ -194,29 +210,16 @@ export class VideoCaptionModel extends SequelizeModel<VideoCaptionModel> {
     return result
   }
 
+  // ---------------------------------------------------------------------------
+
   static getLanguageLabel (language: string) {
     return VIDEO_LANGUAGES[language] || 'Unknown'
   }
 
-  static deleteAllCaptionsOfRemoteVideo (videoId: number, transaction: Transaction) {
-    const query = {
-      where: {
-        videoId
-      },
-      transaction
-    }
-
-    return VideoCaptionModel.destroy(query)
-  }
-
   static generateCaptionName (language: string) {
     return `${buildUUID()}-${language}.vtt`
   }
 
-  isOwned () {
-    return this.Video.remote === false
-  }
-
   // ---------------------------------------------------------------------------
 
   toFormattedJSON (this: MVideoCaptionFormattable): VideoCaption {
@@ -240,6 +243,10 @@ export class VideoCaptionModel extends SequelizeModel<VideoCaptionModel> {
 
   // ---------------------------------------------------------------------------
 
+  isOwned () {
+    return this.Video.remote === false
+  }
+
   getCaptionStaticPath (this: MVideoCaptionLanguageUrl) {
     return join(LAZY_STATIC_PATHS.VIDEO_CAPTIONS, this.filename)
   }
diff --git a/server/core/models/video/video-job-info.ts b/server/core/models/video/video-job-info.ts
index a1e55948b..3a54ca2c7 100644
--- a/server/core/models/video/video-job-info.ts
+++ b/server/core/models/video/video-job-info.ts
@@ -1,10 +1,10 @@
+import { forceNumber } from '@peertube/peertube-core-utils'
 import { Op, QueryTypes, Transaction } from 'sequelize'
 import { AllowNull, BelongsTo, Column, CreatedAt, Default, ForeignKey, IsInt, Table, Unique, UpdatedAt } from 'sequelize-typescript'
-import { forceNumber } from '@peertube/peertube-core-utils'
-import { VideoModel } from './video.js'
 import { SequelizeModel } from '../shared/sequelize-type.js'
+import { VideoModel } from './video.js'
 
-export type VideoJobInfoColumnType = 'pendingMove' | 'pendingTranscode'
+export type VideoJobInfoColumnType = 'pendingMove' | 'pendingTranscode' | 'pendingTranscription'
 
 @Table({
   tableName: 'videoJobInfo',
@@ -39,6 +39,12 @@ export class VideoJobInfoModel extends SequelizeModel<VideoJobInfoModel> {
   @Column
   pendingTranscode: number
 
+  @AllowNull(false)
+  @Default(0)
+  @IsInt
+  @Column
+  pendingTranscription: number
+
   @ForeignKey(() => VideoModel)
   @Unique
   @Column
diff --git a/server/core/types/models/user/user-notification.ts b/server/core/types/models/user/user-notification.ts
index ca19f8f09..83411d3fd 100644
--- a/server/core/types/models/user/user-notification.ts
+++ b/server/core/types/models/user/user-notification.ts
@@ -16,6 +16,7 @@ import { VideoBlacklistModel } from '../../../models/video/video-blacklist.js'
 import { VideoChannelModel } from '../../../models/video/video-channel.js'
 import { VideoCommentModel } from '../../../models/video/video-comment.js'
 import { VideoImportModel } from '../../../models/video/video-import.js'
+import { VideoCaptionModel } from '@server/models/video/video-caption.js'
 
 type Use<K extends keyof UserNotificationModel, M> = PickWith<UserNotificationModel, K, M>
 
@@ -98,13 +99,17 @@ export module UserNotificationIncludes {
 
   export type UserRegistrationInclude =
     Pick<UserRegistrationModel, 'id' | 'username'>
+
+  export type VideoCaptionInclude =
+    Pick<VideoCaptionModel, 'id' | 'language'> &
+    PickWith<VideoCaptionModel, 'Video', VideoInclude>
 }
 
 // ############################################################################
 
 export type MUserNotification =
   Omit<UserNotificationModel, 'User' | 'Video' | 'VideoComment' | 'Abuse' | 'VideoBlacklist' |
-  'VideoImport' | 'Account' | 'ActorFollow' | 'Plugin' | 'Application' | 'UserRegistration'>
+  'VideoImport' | 'Account' | 'ActorFollow' | 'Plugin' | 'Application' | 'UserRegistration' | 'VideoCaption'>
 
 // ############################################################################
 
@@ -119,4 +124,5 @@ export type UserNotificationModelForApi =
   Use<'Plugin', UserNotificationIncludes.PluginInclude> &
   Use<'Application', UserNotificationIncludes.ApplicationInclude> &
   Use<'Account', UserNotificationIncludes.AccountIncludeActor> &
-  Use<'UserRegistration', UserNotificationIncludes.UserRegistrationInclude>
+  Use<'UserRegistration', UserNotificationIncludes.UserRegistrationInclude> &
+  Use<'VideoCaption', UserNotificationIncludes.VideoCaptionInclude>
diff --git a/server/core/types/models/video/video-caption.ts b/server/core/types/models/video/video-caption.ts
index c00226ba5..3472742a9 100644
--- a/server/core/types/models/video/video-caption.ts
+++ b/server/core/types/models/video/video-caption.ts
@@ -16,7 +16,7 @@ export type MVideoCaptionLanguageUrl =
 
 export type MVideoCaptionVideo =
   MVideoCaption &
-  Use<'Video', Pick<MVideo, 'id' | 'remote' | 'uuid'>>
+  Use<'Video', Pick<MVideo, 'id' | 'name' | 'remote' | 'uuid' | 'url' | 'getWatchStaticPath'>>
 
 // ############################################################################
 
diff --git a/server/tsconfig.types.json b/server/tsconfig.types.json
index 696fe7059..444f9b0a1 100644
--- a/server/tsconfig.types.json
+++ b/server/tsconfig.types.json
@@ -13,7 +13,8 @@
     { "path": "../packages/models" },
     { "path": "../packages/node-utils" },
     { "path": "../packages/server-commands" },
-    { "path": "../packages/typescript-utils" }
+    { "path": "../packages/typescript-utils" },
+    { "path": "../packages/transcription" }
   ],
   "exclude": [
     "tests/"
diff --git a/support/doc/api/openapi.yaml b/support/doc/api/openapi.yaml
index ff7312867..62104f7fd 100644
--- a/support/doc/api/openapi.yaml
+++ b/support/doc/api/openapi.yaml
@@ -4127,7 +4127,7 @@ paths:
   /api/v1/videos/{id}/storyboards:
     get:
       summary: List storyboards of a video
-      description: "**PeerTube** >= 6.0"
+      description: "**PeerTube >= 6.0**"
       operationId: listVideoStoryboards
       tags:
         - Video
@@ -4171,6 +4171,32 @@ paths:
                     items:
                       $ref: '#/components/schemas/VideoCaption'
 
+  /api/v1/videos/{id}/captions/generate:
+    post:
+      summary: Generate a video caption
+      operationId: generateVideoCaption
+      description: "**PeerTube >= 6.2** This feature has to be enabled by the administrator"
+      security:
+        - OAuth2:
+          - user
+      tags:
+        - Video Captions
+      requestBody:
+        content:
+          application/json:
+            schema:
+              type: object
+              properties:
+                forceTranscription:
+                  type: boolean
+                  default: false
+      responses:
+        '204':
+          description: successful operation
+        '404':
+          description: video not found
+
+
   /api/v1/videos/{id}/captions/{captionLanguage}:
     put:
       summary: Add or replace a video caption
@@ -4221,7 +4247,7 @@ paths:
   /api/v1/videos/{id}/chapters:
     get:
       summary: Get chapters of a video
-      description: "**PeerTube** >= 6.0"
+      description: "**PeerTube >= 6.0**"
       operationId: getVideoChapters
       tags:
         - Video Chapters
@@ -4237,7 +4263,7 @@ paths:
                 $ref: '#/components/schemas/VideoChapters'
     put:
       summary: Replace video chapters
-      description: "**PeerTube** >= 6.0"
+      description: "**PeerTube >= 6.0**"
       operationId: replaceVideoChapters
       security:
         - OAuth2:
@@ -4270,7 +4296,7 @@ paths:
   /api/v1/videos/{id}/passwords:
     get:
       summary: List video passwords
-      description: "**PeerTube** >= 6.0"
+      description: "**PeerTube >= 6.0**"
       security:
         - OAuth2:
           - user
@@ -4292,7 +4318,7 @@ paths:
           description: video is not password protected
     put:
       summary: Update video passwords
-      description: "**PeerTube** >= 6.0"
+      description: "**PeerTube >= 6.0**"
       security:
         - OAuth2:
           - user
@@ -4317,7 +4343,7 @@ paths:
   /api/v1/videos/{id}/passwords/{videoPasswordId}:
     delete:
       summary: Delete a video password
-      description: "**PeerTube** >= 6.0"
+      description: "**PeerTube >= 6.0**"
       security:
         - OAuth2:
           - user
@@ -8315,7 +8341,7 @@ components:
               nullable: true
               example: |
                	"**[Want to help to translate this video?](https://weblate.framasoft.org/projects/what-is-peertube-video/)**\r\n\r\n
-                **Take back the control of your videos! [#JoinPeertube](https://joinpeertube.org)**\r\n*A decentralized video hosting network, 
+                **Take back the control of your videos! [#JoinPeertube](https://joinpeertube.org)**\r\n*A decentralized video hosting network,
                 based on free/libre software!*\r\n\r\n**Animation Produced by:** [LILA](https://libreart.info) - [ZeMarmot Team](https://film.zemarmot.net)\r\n
                 *Directed by* Aryeom\r\n*Assistant* Jehan\r\n**Licence**: [CC-By-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/)\r\n\r\n
                 **Sponsored by** [Framasoft](https://framasoft.org)\r\n\r\n**Music**: [Red Step Forward](http://play.dogmazic.net/song.php?song_id=52491) - CC-By Ken Bushima\r\n\r\n
@@ -9683,6 +9709,9 @@ components:
         waitTranscoding:
           description: Whether or not we wait transcoding before publish the video
           type: boolean
+        generateTranscription:
+          description: "**PeerTube >= 6.2** If enabled by the admin, automatically generate a subtitle of the video"
+          type: boolean
         support:
           description: A text tell the audience how to support the video creator
           example: Please support our work on https://soutenir.framasoft.org/en/ <3
diff --git a/support/doc/dependencies.md b/support/doc/dependencies.md
index c6d745607..158766c8d 100644
--- a/support/doc/dependencies.md
+++ b/support/doc/dependencies.md
@@ -2,13 +2,15 @@
 
 :warning: **Warning**: dependencies guide is maintained by the community. Some parts may be outdated! :warning:
 
-Main dependencies version supported by PeerTube:
+Main dependencies supported by PeerTube:
 
  * `node` >=18.x
  * `yarn` 1.x (**must not be >=2.x**)
  * `postgres` >=10.x
  * `redis-server` >=6.x
  * `ffmpeg` >=4.3 (using a ffmpeg static build [is not recommended](https://github.com/Chocobozzz/PeerTube/issues/6308))
+ * `python` >=3.x
+ * `pip`
 
 
 _note_: only **LTS** versions of external dependencies are supported. If no LTS version matching the version constraint is available, only **release** versions are supported.
@@ -48,20 +50,12 @@ _note_: only **LTS** versions of external dependencies are supported. If no LTS
 
 5. Install Python:
 
-On Ubuntu <= bionic (18.04 LTS) or Debian <= Buster:
-
-```
-sudo apt update
-sudo apt install python-dev
-python --version # Should be >= 2.x or >= 3.x
-```
-
 On Ubuntu >= focal (20.04 LTS) or Debian >= Bullseye:
 
 ```
 sudo apt update
-sudo apt install python3-dev python-is-python3 # python-is-python2 should also work
-python --version # Should be >= 2.x or >= 3.x
+sudo apt install python3-dev python3-pip python-is-python3
+python --version # Should be >= 3.x
 ```
 
 6. Install common dependencies:
@@ -85,7 +79,7 @@ sudo systemctl start redis postgresql
 1. Run:
 
 ```
-sudo pacman -S nodejs-lts-fermium yarn ffmpeg postgresql openssl redis git wget unzip python base-devel npm nginx
+sudo pacman -S nodejs-lts-fermium yarn ffmpeg postgresql openssl redis git wget unzip python python-pip base-devel npm nginx
 ```
 
 Now that dependencies are installed, before running PeerTube you should start PostgreSQL and Redis:
@@ -214,7 +208,7 @@ sudo dnf update -y
 
 5. Install PostgreSQL and Python3 and other stuff:
 ```
-sudo dnf install -y nginx postgresql postgresql-server postgresql-contrib openssl gcc-c++ make wget redis git python3
+sudo dnf install -y nginx postgresql postgresql-server postgresql-contrib openssl gcc-c++ make wget redis git python3 python3-pip
 sudo ln -s /usr/bin/python3 /usr/bin/python
 sudo PGSETUP_INITDB_OPTIONS='--auth-host=md5' postgresql-setup --initdb --unit postgresql
 sudo systemctl enable --now redis
diff --git a/support/doc/development/tests.md b/support/doc/development/tests.md
index 2e4c6ff6a..544b2ba2d 100644
--- a/support/doc/development/tests.md
+++ b/support/doc/development/tests.md
@@ -36,11 +36,17 @@ Ensure you also have these commands:
 ```bash
 exiftool --help
 parallel --help
+
+# For transcription tests
+whisper --help
+whisper-ctranslate2 --help
+jiwer --help
 ```
 
 Otherwise, install the packages. On Debian-based systems (like Debian, Ubuntu or Mint):
 ```bash
 sudo apt-get install parallel libimage-exiftool-perl
+sudo pip install -r packages/tests/requirements.txt -r packages/transcription-devtools/requirements.txt
 ```
 
 ### Test
diff --git a/tsconfig.eslint.json b/tsconfig.eslint.json
index 61542e14e..9ef5a40b7 100644
--- a/tsconfig.eslint.json
+++ b/tsconfig.eslint.json
@@ -24,7 +24,7 @@
     { "path": "./apps/peertube-cli" },
     { "path": "./packages/core-utils" },
     { "path": "./packages/ffmpeg" },
-    { "path": "./packages/jiwer" },
+    { "path": "./packages/transcription-devtools" },
     { "path": "./packages/models" },
     { "path": "./packages/node-utils" },
     { "path": "./packages/server-commands" },