V2 est le modèle de transcription simplifié, rattaché à l'utilisateur. Il privilégie les hits de cache rapides, des réponses cohérentes et aucune traduction surprise, sauf si vous demandez une langue.
https://www.youtubetranscript.dev/api/v2Collez un lien ou un ID YouTube et recevez la transcription, généralement en quelques secondes.
/transcribeEnvoyez jusqu'à 3 000 vidéos en une requête et soyez notifié à la fin.
/batchTransformez une playlist ou une chaîne en liste de vidéos prête à transcrire.
/playlists/resolveImportez un fichier et transcrivez-le par reconnaissance vocale.
/uploadsListez, récupérez, traduisez ou supprimez vos transcriptions.
/transcriptsSuivez les tâches de reconnaissance vocale ou recevez un webhook à la fin.
/jobs/{job_id}Nouveau : un seul paramètre mode pour toutes les sources de transcription
Choisissez auto, captions_only, human_captions_only, speech_to_text_only au lieu de combiner source et allow_asr. Les requêtes existantes continuent de fonctionner exactement comme avant. Voir comment ça marche
Téléchargez le YAML pour l'importer dans Postman ou Insomnia, générer des clients ou le coller dans ChatGPT / Claude.
Le moyen le plus rapide pour une vidéo à la fois. Les transcriptions basées sur les sous-titres sont renvoyées en quelques secondes ; l'ASR s'exécute de manière asynchrone et s'interroge via /api/v2/jobs/{job_id}.
Transcribe a single YouTube video. Returns immediately for caption-based videos. With mode=auto it falls back to ASR (async) when captions are unavailable: you get status processing and a job_id to poll at /api/v2/jobs/{job_id}, or the result is POSTed to webhook_url if you set one. Responses include the mode the request ran with.
/api/v2/transcribevideostringrequisYouTube URL or 11-character video ID. Required unless upload_id is set.
upload_idstringTranscribe an uploaded media file instead of a YouTube video (see /uploads). Forces the ASR path (mode must be omitted or speech_to_text_only); webhook_url required. The upload id is returned as video_id in results.
modeauto | captions_only | human_captions_only | speech_to_text_onlyWhere the transcript comes from (recommended). auto: captions first, speech-to-text (ASR) only if the video has none. captions_only: any captions, never ASR. human_captions_only: human-made captions only. speech_to_text_only: always transcribe the audio. auto and speech_to_text_only can charge ASR credits. Replaces source + allow_asr; cannot be combined with them. Note: mode=auto is not the same as legacy source=auto (captions only, which is mode=captions_only). When omitted, your account default from the API Dashboard applies: auto for accounts created from October 2026, captions_only for older accounts unless changed. ASR modes work without webhook_url: poll the returned job_id.
languageISO 639-1Preferred transcript language (e.g. en). If unavailable, captions in another language may be returned: transcript.language reports the actual language.
sourceauto | manual | asrLegacy, still supported: use mode instead. Default: auto. asr transcribes the audio (async, requires webhook_url).
allow_asrbooleanLegacy, still supported: use mode=auto instead. Falls back to ASR if captions fail.
formatobject{ timestamp, paragraphs, words }: include extra structure in the transcript.
webhook_urluriWhen set, ASR results are POSTed to your URL. Optional with mode (poll the job_id instead); required for legacy source=asr. Ignored by captions_only and human_captions_only, which never use ASR.
asr_optionsobjectRich ASR config: language (+ codeSwitching, medicalMode, keyTerms), translateTo, speaker diarization (speakerLabels, speakerId + type/known values, multichannel, exact or min/max expected speakers), speech understanding (topics, summary, moderation, PII, profanity, sentiment), and formatting (punctuation, text formatting, filler words, custom date/phone/email patterns). Billable add-ons increase cost: preview with /transcribe/estimate. See the AsrOptions schema in the OpenAPI spec.
curl -X POST https://www.youtubetranscript.dev/api/v2/transcribe \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"video": "dQw4w9WgXcQ",
"mode": "auto",
"language": "en",
"format": { "timestamp": true, "paragraphs": true }
}'{
"request_id": "req_abc123",
"status": "completed",
"data": {
"video_id": "dQw4w9WgXcQ",
"video_title": "Example video",
"transcript": {
"text": "Welcome to this tutorial...",
"language": "en",
"source": "auto",
"segments": [
{ "text": "Welcome to this tutorial.", "start": 0, "end": 2500 }
]
}
},
"credits_used": 1,
"mode": "auto"
}Dry-run credit estimate. Returns the ASR base cost plus any add-on surcharge for the requested asr_options without creating a job or charging credits.
/api/v2/transcribe/estimatevideostringrequisYouTube URL or 11-character video ID.
modeauto | captions_only | human_captions_only | speech_to_text_onlySame values as /transcribe. Use speech_to_text_only (or auto) to preview the ASR cost.
sourceauto | manual | asrLegacy, still supported: use mode instead. Defaults to asr for cost preview.
asr_optionsobjectSame shape as /transcribe. Add-on charges are prorated by duration.
curl -X POST https://www.youtubetranscript.dev/api/v2/transcribe/estimate \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"video": "dQw4w9WgXcQ",
"mode": "speech_to_text_only",
"asr_options": {
"speakerLabels": true,
"summarization": true,
"translateTo": "es"
}
}'{
"request_id": "req_abc123",
"video_id": "dQw4w9WgXcQ",
"source": "asr",
"mode": "speech_to_text_only",
"duration_seconds": 213,
"duration_minutes": 4,
"max_duration_minutes": 480,
"exceeds_max_duration": false,
"base_credits": 3,
"addon_credits": 2,
"estimated_credits": 5
}Poll a single transcription job. Use this for ASR jobs returned from /transcribe. Returns 202 while running and 200 when complete. Completed ASR jobs include asr_intelligence (summary, topics, sentiment, etc.) when those options were enabled.
/api/v2/jobs/{job_id}job_idpathrequisIdentifier returned in the original transcribe response.
include_segmentsbooleanInclude segment-level timestamps in the transcript payload.
include_paragraphsbooleanInclude paragraph groupings.
include_wordsbooleanInclude word-level timestamps (ASR only).
curl -X GET https://www.youtubetranscript.dev/api/v2/jobs/JOB_ID \
-H "Authorization: Bearer YOUR_API_KEY"{
"request_id": "req_abc123",
"status": "completed",
"data": {
"video_id": "dQw4w9WgXcQ",
"transcript": {
"text": "Welcome to this tutorial...",
"language": "en",
"source": "asr",
"segments": [{ "text": "Welcome to this tutorial.", "start": 0, "end": 2500 }]
}
},
"credits_used": 6
}Transcrivez vos propres fichiers audio ou vidéo via le pipeline ASR. Trois étapes : enregistrez l'upload, envoyez les octets en PUT vers l'URL signée, finalisez: puis appelez /transcribe avec upload_id. Les résultats arrivent sur votre webhook et s'interrogent comme n'importe quel job ASR.
Register an audio/video file upload and get a signed URL. PUT the raw file bytes to signed_url (with the file's Content-Type), finalize via /uploads/{upload_id}/complete, then transcribe with upload_id on /transcribe. Max 2GB and 8 hours per file.
/api/v2/uploadsfilenamestringrequisOriginal filename. The extension is preserved for content sniffing.
mime_typestringrequisMust start with audio/ or video/ (e.g. audio/mpeg, video/mp4).
size_bytesintegerrequisFile size in bytes. Max 2147483648 (2GB).
duration_secintegerrequisMedia duration in seconds. Max 28800 (8 hours). Used for the ASR credit estimate.
curl -X POST https://www.youtubetranscript.dev/api/v2/uploads \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"filename": "interview.mp3",
"mime_type": "audio/mpeg",
"size_bytes": 48211234,
"duration_sec": 1820
}'
# Then PUT the raw file bytes to the returned signed_url:
curl -X PUT "SIGNED_URL_FROM_RESPONSE" \
-H "Content-Type: audio/mpeg" \
--data-binary @interview.mp3
# Then finalize:
curl -X POST https://www.youtubetranscript.dev/api/v2/uploads/UPLOAD_ID/complete \
-H "Authorization: Bearer YOUR_API_KEY"{
"upload_id": "up_a1B2c3D4",
"bucket": "asr-media",
"path": "user-uuid/up_a1B2c3D4/interview.mp3",
"token": "eyJhbGciOiJIUzI1NiJ9...",
"signed_url": "https://api.youtubetranscript.dev/storage/v1/object/upload/sign/asr-media/...?token=..."
}Finalize an upload after the PUT to signed_url succeeds. Verifies the file exists in storage and marks it ready. The upload can then be transcribed by passing upload_id to /transcribe (source="asr", webhook_url required).
/api/v2/uploads/{upload_id}/completeupload_idpathrequisUpload id returned by POST /uploads.
curl -X POST https://www.youtubetranscript.dev/api/v2/uploads/up_a1B2c3D4/complete \
-H "Authorization: Bearer YOUR_API_KEY"{
"upload_id": "up_a1B2c3D4",
"status": "uploaded",
"filename": "interview.mp3",
"duration_sec": 1820
}Transcrivez de nombreuses vidéos en une seule requête, puis interrogez le statut ou recevez les résultats via webhook.
Submit up to 3,000 videos in one request. Returns within 2 seconds; poll the returned poll_url or wait for the webhook.
/api/v2/batchvideo_idsstring[]requisUp to 3,000 IDs depending on plan. Free = 10, Basic = 500, Pro = 1,500, Business = 3,000.
languageISO 639-1Preferred language for every video.
modeauto | captions_only | human_captions_only | speech_to_text_onlyWhere the transcript comes from (recommended). auto: captions first, speech-to-text (ASR) only if the video has none. captions_only: any captions, never ASR. human_captions_only: human-made captions only. speech_to_text_only: always transcribe the audio. auto and speech_to_text_only can charge ASR credits. Replaces source + allow_asr; cannot be combined with them. Note: mode=auto is not the same as legacy source=auto (captions only, which is mode=captions_only). For batch, auto and speech_to_text_only require webhook_url. When omitted, your API Dashboard default applies (an ASR default only when webhook_url is set).
sourceauto | manual | asrLegacy, still supported: use mode instead. Default: auto. source=asr requires webhook_url.
allow_asrbooleanLegacy, still supported: use mode=auto instead. webhook_url is required.
formatobject{ timestamp, paragraphs, words }: applied to every result.
webhook_urluriDelivers the completed batch payload. Required for mode=auto or speech_to_text_only (legacy: allow_asr=true or source=asr).
asr_optionsobjectRich ASR config applied to every video transcribed via ASR (same shape as /transcribe). Billable add-ons apply per video.
curl -X POST https://www.youtubetranscript.dev/api/v2/batch \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"video_ids": ["dQw4w9WgXcQ", "jNQXAC9IVRw"],
"mode": "auto",
"webhook_url": "https://your-domain.com/webhook",
"format": { "timestamp": true }
}'{
"batch_id": "550e8400-e29b-41d4-a716-446655440000",
"status": "processing",
"results": [],
"summary": { "total": 2, "succeeded": 0, "failed": 0, "processing": 2 },
"poll_url": "https://www.youtubetranscript.dev/api/v2/batch/550e8400-e29b-41d4-a716-446655440000",
"created_at": "2026-05-15T10:00:00Z"
}Check the status of a batch. Returns 202 while processing and 200 when results are ready. Stop polling once status is "completed", "partial", or "failed": finished batches never change, and re-fetching one more than 60 times in 24h returns 429 with a Retry-After header.
/api/v2/batch/{batch_id}batch_idpathrequisBatch identifier returned from POST /batch.
curl -X GET https://www.youtubetranscript.dev/api/v2/batch/BATCH_ID \
-H "Authorization: Bearer YOUR_API_KEY"{
"batch_id": "550e8400-e29b-41d4-a716-446655440000",
"status": "completed",
"summary": { "total": 2, "succeeded": 2, "failed": 0, "processing": 0 },
"results": [
{ "request_id": "req_1", "status": "completed", "data": { "video_id": "dQw4w9WgXcQ" } },
{ "request_id": "req_2", "status": "completed", "data": { "video_id": "jNQXAC9IVRw" } }
],
"credits_used": 2,
"webhook_delivered": true,
"completed_at": "2026-05-15T10:04:00Z"
}Listez, récupérez, traduisez et supprimez vos transcriptions enregistrées, et listez les jobs asynchrones. Tous les endpoints acceptent la même authentification par clé API.
List your transcription jobs, newest first. Filter by status and paginate. Use the returned job_id with GET /api/v2/jobs/{job_id} to fetch a transcript.
/api/v2/jobspagequeryPage number (default 1).
limitqueryItems per page (default 50, max 100).
statusquerycompleted | processing | failed | requires_asr_confirmation.
curl "https://www.youtubetranscript.dev/api/v2/jobs?status=processing&limit=25" \
-H "Authorization: Bearer YOUR_API_KEY"{
"jobs": [
{
"job_id": "9c2...",
"video_id": "dQw4w9WgXcQ",
"video_title": "Example video",
"language": "en",
"source": "asr",
"status": "processing",
"error": null,
"credits_used": 0,
"created_at": "2026-07-06T12:00:00Z",
"started_at": "2026-07-06T12:00:01Z",
"completed_at": null
}
],
"count": 1,
"pagination": { "page": 1, "limit": 25, "total": 1, "totalPages": 1, "hasMore": false }
}List stored transcripts with search, status, language, and date filters. Metadata only (no full text) unless include_segments=true.
/api/v2/transcriptspagequeryPage number (default 1).
limitqueryItems per page (default 50, max 100).
searchqueryMatch video id, title, or transcript text.
statusqueryFilter by status; omit for succeeded.
languagequeryFilter by language code.
date_fromqueryISO date-time lower bound.
date_toqueryISO date-time upper bound.
include_segmentsqueryInclude segments in each item.
curl "https://www.youtubetranscript.dev/api/v2/transcripts?page=1&limit=50&status=succeeded" \
-H "Authorization: Bearer YOUR_API_KEY"{
"history": [
{
"id": "b1f...",
"video_id": "dQw4w9WgXcQ",
"video_title": "Example video",
"language": "en",
"source_kind": "auto",
"status": "succeeded",
"created_at": "2026-07-06T12:00:00Z",
"word_count": 1234,
"segment_count": 88,
"credits_used": 1
}
],
"count": 1,
"pagination": { "page": 1, "limit": 50, "total": 1, "totalPages": 1, "hasMore": false }
}Fetch the best owned transcript for a video, including asr_intelligence (summary, topics, sentiment, etc.), plus summary and mind_map when generated.
/api/v2/transcripts/{video_id}video_idpathrequisYouTube 11-character video ID.
idquerySpecific transcript/job id.
languagequeryPreferred language version.
sourcequeryauto | manual | asr.
include_timestampsqueryInclude segments (default true).
curl "https://www.youtubetranscript.dev/api/v2/transcripts/dQw4w9WgXcQ?language=en" \
-H "Authorization: Bearer YOUR_API_KEY"{
"id": "b1f...",
"video_id": "dQw4w9WgXcQ",
"language": "en",
"text": "Welcome to this tutorial...",
"segments": [{ "text": "Welcome to this tutorial.", "start": 0, "end": 2500 }],
"has_timestamps": true,
"source_kind": "asr",
"video_title": "Example video",
"summary": null,
"mind_map": null,
"asr_intelligence": {
"summary": "A short overview of the video...",
"topics": [{ "text": "Technology", "relevance": 0.92 }],
"sentiment": [{ "text": "Great tutorial", "sentiment": "POSITIVE" }]
}
}AI key moments, themes & keywords for a stored transcript: a short TL;DR, the key moments worth finding again (each a short label plus the exact quote, with where it starts and ends), its main themes, and each keyword with the timestamps where it's said. 1 credit per transcript, charged once: calling it again (or GET) is free. Waits up to ~25s; a 202 "processing" means it's paid for and still running, so poll GET. If detection fails the credit is refunded. Transcripts too short to analyse return empty themes for 0 credits. 402 without enough credits.
/api/v2/transcripts/{video_id}/themesvideo_idpathrequisYouTube 11-character video ID of a transcript you own.
idquerySpecific transcript/job id.
languagequeryWhich language version to analyse.
sourcequeryauto | manual | asr.
curl -X POST "https://www.youtubetranscript.dev/api/v2/transcripts/dQw4w9WgXcQ/themes?language=en" \
-H "Authorization: Bearer YOUR_API_KEY"{
"video_id": "dQw4w9WgXcQ",
"language": "en",
"status": "completed",
"credits_used": 1,
"credits_remaining": 41,
"tldr": "A promise of lifelong loyalty, sung as a list of things he'll never do.",
"moments": [
{
"id": "m1",
"theme_id": "t1",
"label": "The six promises of the chorus",
"quote": "Never gonna give you up, never gonna let you down",
"color_hex": "#0d9488",
"start": 43000,
"end": 47500,
"segment_start": 4,
"segment_end": 5
}
],
"themes": [
{
"id": "t1",
"name": "Commitment",
"summary": "Promises never to leave or let the listener down.",
"color": "teal",
"color_hex": "#0d9488",
"share": 0.62,
"mention_count": 18,
"moment_count": 3,
"keywords": [
{
"id": "k1",
"term": "never gonna give you up",
"aliases": [],
"definition": "The song's central promise.",
"mention_count": 6,
"mentions": [{ "segment_index": 4, "start": 43000 }]
}
]
}
]
}Read key moments, themes & keywords you've already unlocked. Never charges and never starts the AI: returns status "locked" (with credits_required) until you POST, then "completed" with the same body as POST, or 202 "processing" while detection is still running.
/api/v2/transcripts/{video_id}/themesvideo_idpathrequisYouTube 11-character video ID of a transcript you own.
idquerySpecific transcript/job id.
languagequeryWhich language version to analyse.
sourcequeryauto | manual | asr.
curl "https://www.youtubetranscript.dev/api/v2/transcripts/dQw4w9WgXcQ/themes?language=en" \
-H "Authorization: Bearer YOUR_API_KEY"{
"video_id": "dQw4w9WgXcQ",
"language": "en",
"status": "locked",
"credits_required": 1
}Translate an owned transcript into a target language using YouTube captions or the AI provider. Returns free if you already own the translation.
/api/v2/transcripts/{video_id}/translatevideo_idpathrequisYouTube 11-character video ID.
target_languageISO 639-1requisLanguage to translate into.
source_languageISO 639-1Source transcript language to translate from.
providerai | youtubeTranslation provider.
allow_ai_fallbackbooleanFall back to AI when YouTube has no translation.
curl -X POST "https://www.youtubetranscript.dev/api/v2/transcripts/dQw4w9WgXcQ/translate" \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{ "target_language": "es", "provider": "ai" }'{
"success": true,
"message": "Translation completed",
"language": "es",
"credits_used": 4,
"provider_used": "ai"
}List the languages you already own for a video plus the YouTube translation targets available for it.
/api/v2/transcripts/{video_id}/languagesvideo_idpathrequisYouTube 11-character video ID.
include_youtube_defaultqueryInclude YouTube translation targets (default true).
curl "https://www.youtubetranscript.dev/api/v2/transcripts/dQw4w9WgXcQ/languages" \
-H "Authorization: Bearer YOUR_API_KEY"{
"video_id": "dQw4w9WgXcQ",
"languages": ["en", "es"],
"youtube_translation_languages": ["fr", "de", "ja"]
}Delete owned transcript(s) for a video. Optional language/source query params narrow the deletion; otherwise every language/source is removed. Also deletes bulk by id via POST /api/v2/transcripts/bulk-delete.
/api/v2/transcripts/{video_id}video_idpathrequisYouTube 11-character video ID.
languagequeryOnly delete this language.
sourcequeryOnly delete this source (auto | manual | asr).
curl -X DELETE "https://www.youtubetranscript.dev/api/v2/transcripts/dQw4w9WgXcQ" \
-H "Authorization: Bearer YOUR_API_KEY"{
"success": true,
"deleted": 2,
"message": "Successfully deleted 2 transcripts"
}Résolvez une playlist en IDs de vidéos, transcrivez-les avec /api/v2/batch, puis consultez le lot de la playlist et son historique.
Resolve a YouTube playlist by URL or ID into its video list (metadata only). Pair the returned video_id values with /api/v2/batch to transcribe them.
/api/v2/playlists/resolvePlans payantsplaylist_urlstringPlaylist URL. One of playlist_url or playlist_id is required.
playlist_idstringPlaylist ID (the list= value).
limitintegerMax videos returned (1-100).
curl -X POST https://www.youtubetranscript.dev/api/v2/playlists/resolve \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{ "playlist_url": "https://www.youtube.com/playlist?list=PLxxxx", "limit": 100 }'{
"playlist_id": "PLxxxx",
"title": "My playlist",
"total": 42,
"truncated": false,
"items": [
{ "video_id": "dQw4w9WgXcQ", "title": "Example video" }
]
}Fetch a playlist batch and per-video transcription job status. For overall batch progress, poll GET /api/v2/batch/{batch_id}.
/api/v2/playlists/{playlist_id}playlist_idpathrequisBatch ID returned from /playlists/resolve.
curl "https://www.youtubetranscript.dev/api/v2/playlists/{playlist_batch_id}" \
-H "Authorization: Bearer YOUR_API_KEY"{
"playlist": { "id": "3f1...", "playlist_id": "PLxxxx", "title": "My playlist" },
"videos": [
{ "video_id": "dQw4w9WgXcQ", "video_title": "Example video", "status": "succeeded" }
]
}List recent playlist batches for the authenticated user.
/api/v2/playlists/historycurl "https://www.youtubetranscript.dev/api/v2/playlists/history" \
-H "Authorization: Bearer YOUR_API_KEY"{
"history": [
{ "id": "3f1...", "playlist_id": "PLxxxx", "title": "My playlist", "total_videos": 42 }
],
"count": 1
}Flux de transcription d'une chaîne :1. POST /api/v2/channels/resolve: récupère les vidéos de la chaîne (métadonnées uniquement).2. POST /api/v2/batch: envoyez les valeurs video_id renvoyées pour lancer la transcription. Utilisez un webhook_url pour recevoir les résultats une fois terminé.3. GET /api/v2/batch/{batch_id}: interrogez le lot pour suivre la progression globale et obtenir les résultats agrégés.4. GET /api/v2/channels/{channel_id}: récupère un lot de chaîne avec le statut du job de chaque vidéo ; /channels/history liste les lots de chaînes précédents.
Fetch the uploads of a YouTube channel by URL, channel ID, or @handle. Returns video metadata only: pair the returned video_id values with /api/v2/batch to transcribe them.
/api/v2/channels/resolvePlans payantschannel_urlstringrequisYouTube channel URL, channel ID, or @handle. One of channel_url, channel_id, or handle must be provided. Handles are the public ID without the @ (for https://www.youtube.com/@jawed, the ID is jawed).
limitintegerMaximum uploads returned. Defaults to the plan maximum.
Per-request cap: Basic 500 videos · Pro 1,500 videos · Business 3,000 videos. Monthly job cap: Basic 3 · Pro 8 · Business unlimited.
Plan-based: Free 1 req/s · Basic 25 req/s · Pro 50 req/s · Business 100 req/s. Exceeding the limit returns 429 with a Retry-After header.
curl -X POST https://www.youtubetranscript.dev/api/v2/channels/resolve \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"handle": "@youtube",
"limit": 50
}'{
"channel_id": "UC_x5XG1OV2P6uZZ5FSM9Ttw",
"channel_title": "YouTube",
"subscriber_count": 42000000,
"video_count": 1200,
"total": 1,
"truncated": false,
"items": [
{
"video_id": "dQw4w9WgXcQ",
"title": "Latest upload",
"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
"published_at": "2026-05-01T12:00:00Z",
"duration_seconds": 420
}
]
}Fetch a channel batch and the per-video transcription job status. For overall batch status (counts + progress), poll GET /api/v2/batch/{batch_id} instead: use this endpoint when you also need each video's individual job state.
/api/v2/channels/{channel_id}channel_idpathrequisBatch ID returned from /channels/resolve.
curl -X GET https://www.youtubetranscript.dev/api/v2/channels/CHANNEL_BATCH_ID \
-H "Authorization: Bearer YOUR_API_KEY"{
"channel": {
"id": "550e8400-e29b-41d4-a716-446655440000",
"channel_id": "UC_x5XG1OV2P6uZZ5FSM9Ttw",
"channel_title": "YouTube",
"total_videos": 2,
"completed_count": 1,
"failed_count": 0,
"status": "processing",
"created_at": "2026-05-15T10:00:00Z"
},
"videos": [
{
"id": "job_123",
"video_id": "dQw4w9WgXcQ",
"video_title": "Latest upload",
"status": "succeeded",
"created_at": "2026-05-15T10:00:00Z",
"finished_at": "2026-05-15T10:00:10Z"
}
]
}List recent channel batches for the authenticated user.
/api/v2/channels/historylimitqueryMax items (default 50, max 100).
curl -X GET "https://www.youtubetranscript.dev/api/v2/channels/history?limit=50" \
-H "Authorization: Bearer YOUR_API_KEY"{
"history": [
{
"id": "550e8400-e29b-41d4-a716-446655440000",
"channel_id": "UC_x5XG1OV2P6uZZ5FSM9Ttw",
"channel_title": "YouTube",
"total_videos": 25,
"completed_count": 25,
"failed_count": 0,
"status": "completed",
"created_at": "2026-05-15T10:00:00Z",
"completed_at": "2026-05-15T10:12:00Z"
}
],
"count": 1
}Newest uploads from the channels you transcribe most, excluding videos you have already transcribed. Results are interleaved across channels, newest first. Free: no credits are charged. Pass the returned url or video_id to /api/v2/transcribe or /api/v2/batch.
/api/v2/recommendations/videoslimitqueryMax videos returned (default 6, max 30).
channelsqueryHow many of your top channels to draw from (default 3, max 10).
curl -X GET "https://www.youtubetranscript.dev/api/v2/recommendations/videos?limit=6" \
-H "Authorization: Bearer YOUR_API_KEY"{
"recommendations": [
{
"video_id": "dQw4w9WgXcQ",
"title": "Latest upload",
"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
"thumbnail_url": "https://i.ytimg.com/vi/dQw4w9WgXcQ/mqdefault.jpg",
"published_at": "2026-10-17T15:00:00Z",
"duration_seconds": 1089,
"channel_id": "UC_x5XG1OV2P6uZZ5FSM9Ttw",
"channel_title": "YouTube"
}
],
"channels": [
{
"channel_id": "UC_x5XG1OV2P6uZZ5FSM9Ttw",
"channel_title": "YouTube",
"transcribed_count": 12
}
],
"count": 1
}Rechercher, puis transcrire :1. GET /api/v2/search/videos (/shorts, /channels, /playlists) : trouvez du contenu par mot-clé. 1 crédit par requête.2. Envoyez l'url d'une vidéo à POST /api/v2/transcribe, ou l'url d'une chaîne ou d'une playlist à /api/v2/channels/resolve ou /api/v2/playlists/resolve.3. Passez next_page_token comme page_token pour obtenir plus de résultats (1 crédit par page).
Search YouTube videos by keyword. Live streams are included with type "live". Pass a result's url to /api/v2/transcribe to get its transcript. 1 credit per request, including each next page. Refunded if the search fails.
/api/v2/search/videosqqueryrequisSearch text, up to 200 characters.
durationqueryVideo length: under_3_min, 3_to_20_min or over_20_min.
upload_datequeryOnly results uploaded within: today, this_week, this_month or this_year.
sort_byqueryrelevance (default) or popular.
regionqueryTwo-letter country code to localize results, e.g. US.
page_tokenqueryThe next_page_token from a previous response, to get the next page. null means there are no more results.
Plan-based: Free 1 req/s · Basic 25 req/s · Pro 50 req/s · Business 100 req/s. Exceeding the limit returns 429 with a Retry-After header.
curl "https://www.youtubetranscript.dev/api/v2/search/videos?q=python%20tutorial&duration=over_20_min" \
-H "Authorization: Bearer YOUR_API_KEY"{
"request_id": "6f1c2a9e-4b7d-4f2a-9c1e-2d3b4a5c6d7e",
"type": "videos",
"query": "python tutorial",
"results": [
{
"type": "video",
"video_id": "kqtD5dpn9C8",
"url": "https://www.youtube.com/watch?v=kqtD5dpn9C8",
"title": "Python for Beginners - Learn Coding with Python in 1 Hour",
"thumbnail": "https://i.ytimg.com/vi/kqtD5dpn9C8/hq720.jpg",
"channel": {
"channel_id": "UCWv7vMbMWH4-V0ZXdmDpPBA",
"title": "Programming with Mosh",
"handle": "programmingwithmosh",
"url": "https://www.youtube.com/@programmingwithmosh",
"thumbnail": "https://yt3.ggpht.com/..."
},
"view_count": 25574982,
"view_count_text": "25,574,982 views",
"published_at": null,
"published_text": "6 years ago",
"duration_seconds": 3606,
"duration_text": "1:00:06",
"badges": []
}
],
"next_page_token": "cDA6RWdJUUFV...",
"credits_used": 1,
"credits_remaining": 4602
}Search YouTube Shorts by keyword. 1 credit per request, including each next page. Refunded if the search fails.
/api/v2/search/shortsqqueryrequisSearch text, up to 200 characters.
upload_datequeryOnly results uploaded within: today, this_week, this_month or this_year.
sort_byqueryrelevance (default) or popular.
regionqueryTwo-letter country code to localize results, e.g. US.
page_tokenqueryThe next_page_token from a previous response, to get the next page. null means there are no more results.
Plan-based: Free 1 req/s · Basic 25 req/s · Pro 50 req/s · Business 100 req/s. Exceeding the limit returns 429 with a Retry-After header.
curl "https://www.youtubetranscript.dev/api/v2/search/shorts?q=python%20tips" \
-H "Authorization: Bearer YOUR_API_KEY"{
"request_id": "0b8e7d6c-5a4f-4e3d-8c2b-1a0f9e8d7c6b",
"type": "shorts",
"query": "python tips",
"results": [
{
"type": "short",
"video_id": "gbYLsUngzLc",
"url": "https://www.youtube.com/watch?v=gbYLsUngzLc",
"title": "Top 10 Most Important Python Functions You Must Know!",
"thumbnail": "https://i.ytimg.com/vi/gbYLsUngzLc/oar2.jpg",
"channel": null,
"view_count": 1200000,
"view_count_text": "1.2M views",
"published_at": null,
"published_text": null,
"duration_seconds": null,
"duration_text": null,
"badges": []
}
],
"next_page_token": "cDA6RWdJUUNR...",
"credits_used": 1,
"credits_remaining": 4601
}Search YouTube channels by keyword. Pass a result's url to /api/v2/channels/resolve to list its videos. 1 credit per request, including each next page. Refunded if the search fails.
/api/v2/search/channelsqqueryrequisSearch text, up to 200 characters.
upload_datequeryOnly results uploaded within: today, this_week, this_month or this_year.
sort_byqueryrelevance (default) or popular.
regionqueryTwo-letter country code to localize results, e.g. US.
page_tokenqueryThe next_page_token from a previous response, to get the next page. null means there are no more results.
Plan-based: Free 1 req/s · Basic 25 req/s · Pro 50 req/s · Business 100 req/s. Exceeding the limit returns 429 with a Retry-After header.
curl "https://www.youtubetranscript.dev/api/v2/search/channels?q=python" \
-H "Authorization: Bearer YOUR_API_KEY"{
"request_id": "9a8b7c6d-5e4f-4a3b-2c1d-0e9f8a7b6c5d",
"type": "channels",
"query": "python",
"results": [
{
"type": "channel",
"channel_id": "UCKQdc0-Targ4nDIAUrlfKiA",
"url": "https://www.youtube.com/@PythonSimplified",
"title": "Python Simplified",
"handle": "PythonSimplified",
"description": "Hi everyone! My name is Mariya and I'm a software developer...",
"thumbnail": "https://yt3.googleusercontent.com/...",
"subscriber_count": 286000,
"subscriber_count_text": "286K",
"badges": ["Verified"]
}
],
"next_page_token": "cDA6RWdJUUFn...",
"credits_used": 1,
"credits_remaining": 4600
}Search YouTube playlists by keyword. Pass a result's url to /api/v2/playlists/resolve to list its videos. 1 credit per request, including each next page. Refunded if the search fails.
/api/v2/search/playlistsqqueryrequisSearch text, up to 200 characters.
upload_datequeryOnly results uploaded within: today, this_week, this_month or this_year.
sort_byqueryrelevance (default) or popular.
regionqueryTwo-letter country code to localize results, e.g. US.
page_tokenqueryThe next_page_token from a previous response, to get the next page. null means there are no more results.
Plan-based: Free 1 req/s · Basic 25 req/s · Pro 50 req/s · Business 100 req/s. Exceeding the limit returns 429 with a Retry-After header.
curl "https://www.youtubetranscript.dev/api/v2/search/playlists?q=python%20course" \
-H "Authorization: Bearer YOUR_API_KEY"{
"request_id": "1c2d3e4f-5a6b-4c7d-8e9f-0a1b2c3d4e5f",
"type": "playlists",
"query": "python course",
"results": [
{
"type": "playlist",
"playlist_id": "PLsyeobzWxl7omDoEYrrf3oXvXxa6MPgek",
"url": "https://www.youtube.com/playlist?list=PLsyeobzWxl7omDoEYrrf3oXvXxa6MPgek",
"title": "Python Tutorial | Complete Python Course (Beginner to Advanced)",
"thumbnail": "https://i.ytimg.com/vi/YZkyL-f-YXY/hq720.jpg",
"video_count": 54,
"first_video_id": "YZkyL-f-YXY",
"channel": {
"channel_id": "UC59K-uG2A5ogwIrHw4bmlEg",
"title": "Telusko",
"handle": "Telusko",
"url": "https://www.youtube.com/@Telusko",
"thumbnail": null
}
}
],
"next_page_token": "cDA6RWdJUUF3...",
"credits_used": 1,
"credits_remaining": 4599
}SDK officiels, un serveur MCP pour Claude et Cursor, un GPT personnalisé et des modèles prêts à l'emploi pour Make et n8n.
Ajoutez le serveur MCP YouTube Transcript à Claude Desktop, Cursor ou tout client compatible MCP. Il fournit des outils pour transcrire des vidéos, récupérer des transcriptions existantes et consulter votre utilisation.
claude mcp add --transport http youtubetranscript https://mcp.youtubetranscript.dev --header "x-api-token: YOUR_API_KEY"Types TypeScript de premier ordre, réessais, helpers de pagination et vérification des signatures de webhook.
npm install youtube-transcript-apiClients synchrones et asynchrones, type stubs et helpers pour les flux de lots et de webhooks.
pip install youtubetranscriptdevapiAjoutez notre GPT personnalisé dans ChatGPT pour résumer, citer, traduire ou interroger n'importe quelle vidéo YouTube en langage naturel. Propulsé par /api/v2/gpt/* en coulisses.
Importez le modèle (ou copiez l'extrait) pour ajouter la transcription à vos automatisations existantes: flux Notion, Sheets, Slack, Airtable ou de type Zapier.
Module: HTTP > Make a request
- URL: https://www.youtubetranscript.dev/api/v2/transcribe
- Method: POST
- Headers:
Authorization: Bearer {{connection.api_key}}
Content-Type: application/json
- Body type: Raw / JSON
- Request content:
{
"video": "{{1.video_url}}",
"format": { "timestamp": true }
}{
"name": "Transcribe Video",
"type": "n8n-nodes-base.httpRequest",
"parameters": {
"method": "POST",
"url": "https://www.youtubetranscript.dev/api/v2/transcribe",
"authentication": "genericCredentialType",
"genericAuthType": "httpHeaderAuth",
"sendBody": true,
"specifyBody": "json",
"jsonBody": "={ \"video\": $json.video_url }"
}
}Importez la spécification OpenAPI dans Postman, Insomnia ou tout outil compatible OpenAPI 3.
Les erreurs renvoient un corps JSON avec code et message. Le statut HTTP reflète la catégorie.
| Statut HTTP | Code d'erreur | Description |
|---|---|---|
| 400 | invalid_request | Invalid JSON or missing required fields |
| 401 | invalid_api_key | Missing or invalid API key |
| 402 | payment_required | Insufficient credits |
| 404 | no_captions | No captions available and ASR not used |
| 429 | rate_limit_exceeded | Too many requests, check Retry-After |
| 500 | internal_error | Server error, retry with backoff |
Plafonds par plan pour les requêtes soutenues et la taille des lots. Un dépassement de la limite renvoie 429 avec un en-tête Retry-After.
| Plan | Requêtes | Taille de lot |
|---|---|---|
| Free | 1 req/s | 10 videos |
| Basic | 25 req/s | 500 videos |
| Pro | 50 req/s | 1,500 videos |
| Business | 100 req/s | 3,000 videos |
Besoin d'aide pour livrer ? Contactez-nous: nous répondons généralement en quelques heures.