YouTube Transcript Extractor - Timestamps

Extract complete, timestamped captions from up to 5,000 public YouTube videos. Get full transcript text, timed segments and optional SRT subtitles in one Dataset item per video, with actual language, caption source and a clear success or error status.

What this Actor does

Submit public YouTube video URLs and choose one of 47 requested languages. The Actor preserves every text-bearing segment of the selected caption track, joins those segments into plain text, and includes available video and channel metadata. It does not summarize or shorten the transcript.

Manual captions are preferred. When allowed, automatic captions or YouTube-provided translations can be used; isAutoGenerated and isTranslated distinguish these cases. A missing or inaccessible track produces an explicit error item, not a partial successful transcript.

Use cases

• Build searchable video libraries and retrieval-augmented generation (RAG) datasets.
• Process podcasts, lectures and tutorials with timestamps for source attribution.
• Obtain SRT text for subtitle workflows without downloading media.
• Filter caption sources and detect changes using language metadata and a transcript hash.

How to use

1. Open the Actor in Apify Console.
2. Enter public video URLs and select the transcript language.
3. Optionally require manual captions or enable SRT output.
4. Click Start and open the Dataset. Choose the overview, transcript, RAG or subtitles view.

Input

Field	Meaning
videoUrls	Required: 1–5,000 public watch, short-link, Shorts, live or embed URLs. Duplicate video IDs are processed once, preserving first-seen order.
language	English (en) by default. Choose from the 47-language dropdown. A base code such as es can match a regional variant; a regional request remains specific. The returned languageUsed identifies the actual track.
captionPolicy	any (default) prefers manual captions, then permits automatic or translated tracks. manual_only requires a manually provided track in the requested language; it does not substitute automatic captions or generated translations.
includeSrt	false by default. Set to true to include SRT text derived from every returned segment, at no additional transcript event charge.

Unsupported language codes and invalid input are rejected. The Actor does not silently return an unrelated language. This example enables SRT:

{
  "includeSrt": true,
  "language": "en",
  "captionPolicy": "any",
  "videoUrls": [
    "https://www.youtube.com/watch?v=jNQXAC9IVRw"
  ]
}

Output

One item is written per unique video as processing finishes. Existing segments remain available, and transcript is the complete ordered text joined with spaces.

Fields	Interpretation
videoId, videoUrl, title	Video identity, canonical URL and public title when available.
segments, segmentCount	Complete selected caption track with start, duration and text per segment.
transcript, wordCount, characterCount	Plain text, whitespace-separated token count and Unicode character count. The word count is not a linguistic tokenizer, especially for languages without spaces.
transcriptHash	SHA-256 of the UTF-8 plain text, useful for detecting text changes; it does not include timing or metadata.
languageRequested, languageUsed	Requested language and actual selected track language.
isAutoGenerated, isTranslated	Whether the source uses automatic speech recognition and whether YouTube translated the selected track.
availableLanguages	Distinct language codes advertised by the source, including translation targets. Availability does not guarantee that a later download will succeed. The list may include codes outside the input dropdown.
durationSeconds, channelName, channelId	Public video and channel metadata when exposed.
publishedAt, uploadDate, thumbnailUrl	Source publication timestamp, upload date and image URL. Unknown values are null; a date alone is not turned into an invented timestamp.
srt, exportWarnings	Optional complete SRT string. If any cue lacks a usable end time, SRT is null with SRT_UNAVAILABLE_UNKNOWN_END; the full transcript remains available.
status, errorCode, errorMessage, attemptCount	Result status and number of extraction attempts: normally 1, at most 2, or 0 when the run budget prevents processing.

Use overview for monitoring, transcript for the complete segment array, rag for plain text and provenance, and subtitles for SRT. JSON preserves nested arrays and multiline SRT most naturally; Apify also supports CSV, Excel and other Dataset exports. SRT is a field in the item, not a separate downloadable file.

This is the complete Dataset item from a real successful run:

{
  "videoId": "jNQXAC9IVRw",
  "videoUrl": "https://www.youtube.com/watch?v=jNQXAC9IVRw",
  "status": "success",
  "title": "Me at the zoo",
  "languageRequested": "en",
  "segmentCount": 6,
  "segments": [
    {
      "start": 1.2,
      "duration": 2.16,
      "text": "All right, so here we are, in front of the elephants"
    },
    {
      "start": 5.318,
      "duration": 2.656,
      "text": "the cool thing about these guys is that they have really..."
    },
    {
      "start": 7.974,
      "duration": 4.642,
      "text": "really really long trunks"
    },
    {
      "start": 12.616,
      "duration": 1.751,
      "text": "and that's cool"
    },
    {
      "start": 14.421,
      "duration": 1.312,
      "text": "(baaaaaaaaaaahhh!!)"
    },
    {
      "start": 16.881,
      "duration": 2,
      "text": "and that's pretty much all there is to say"
    }
  ],
  "errorCode": null,
  "errorMessage": null,
  "languageUsed": "en",
  "isAutoGenerated": false,
  "isTranslated": false,
  "availableLanguages": [
    "aa",
    "ab",
    "af",
    "ak",
    "am",
    "ar",
    "as",
    "ay",
    "az",
    "ba",
    "be",
    "bg",
    "bho",
    "bn",
    "bo",
    "br",
    "bs",
    "ca",
    "ceb",
    "co",
    "crs",
    "cs",
    "cy",
    "da",
    "de",
    "dv",
    "dz",
    "ee",
    "el",
    "en",
    "eo",
    "es",
    "et",
    "eu",
    "fa",
    "fi",
    "fil",
    "fj",
    "fo",
    "fr",
    "fy",
    "ga",
    "gaa",
    "gd",
    "gl",
    "gn",
    "gu",
    "gv",
    "ha",
    "haw",
    "he",
    "hi",
    "hmn",
    "hr",
    "ht",
    "hu",
    "hy",
    "id",
    "ig",
    "is",
    "it",
    "iu",
    "ja",
    "jv",
    "ka",
    "kha",
    "kk",
    "kl",
    "km",
    "kn",
    "ko",
    "kri",
    "ku",
    "ky",
    "la",
    "lb",
    "lg",
    "ln",
    "lo",
    "lt",
    "lua",
    "luo",
    "lv",
    "mfe",
    "mg",
    "mi",
    "mk",
    "ml",
    "mn",
    "mr",
    "ms",
    "mt",
    "my",
    "ne",
    "new",
    "nl",
    "no",
    "nso",
    "ny",
    "oc",
    "om",
    "or",
    "os",
    "pa",
    "pam",
    "pl",
    "ps",
    "pt",
    "pt-PT",
    "qu",
    "rn",
    "ro",
    "ru",
    "rw",
    "sa",
    "sd",
    "sg",
    "si",
    "sk",
    "sl",
    "sm",
    "sn",
    "so",
    "sq",
    "sr",
    "ss",
    "st",
    "su",
    "sv",
    "sw",
    "ta",
    "te",
    "tg",
    "th",
    "ti",
    "tk",
    "tn",
    "to",
    "tr",
    "ts",
    "tt",
    "tum",
    "ug",
    "uk",
    "ur",
    "uz",
    "ve",
    "vi",
    "war",
    "wo",
    "xh",
    "yi",
    "yo",
    "zh-Hans",
    "zh-Hant",
    "zu"
  ],
  "durationSeconds": 19,
  "channelName": "jawed",
  "channelId": "UC4QobU6STFB0P71PMvOGN5A",
  "publishedAt": "2005-04-24T03:31:52Z",
  "uploadDate": "2005-04-24",
  "thumbnailUrl": "https://i.ytimg.com/vi/jNQXAC9IVRw/hqdefault.jpg?sqp=-oaymwEmCOADEOgC8quKqQMa8AEB-AG-AoAC8AGKAgwIABABGFUgWShlMA8=&rs=AOn4CLA9eLBatYv9WbkD4BbZ2Im-biSPTw",
  "transcript": "All right, so here we are, in front of the elephants the cool thing about these guys is that they have really... really really long trunks and that's cool (baaaaaaaaaaahhh!!) and that's pretty much all there is to say",
  "wordCount": 39,
  "characterCount": 217,
  "transcriptHash": "083a5f4a06797b6eca382a74707c681a5094f25129bc8447dd0c3557f24ac0cf",
  "srt": "1\n00:00:01,200 --> 00:00:03,360\nAll right, so here we are, in front of the elephants\n\n2\n00:00:05,318 --> 00:00:07,974\nthe cool thing about these guys is that they have really...\n\n3\n00:00:07,974 --> 00:00:12,616\nreally really long trunks\n\n4\n00:00:12,616 --> 00:00:14,367\nand that's cool\n\n5\n00:00:14,421 --> 00:00:15,733\n(baaaaaaaaaaahhh!!)\n\n6\n00:00:16,881 --> 00:00:18,881\nand that's pretty much all there is to say\n",
  "exportWarnings": [],
  "attemptCount": 1
}

Error items have empty segments and transcript, zero counts, and null hash and SRT. Available metadata can remain present. Common codes include NO_TRANSCRIPT, MUSIC_ONLY_CAPTIONS, LANGUAGE_UNAVAILABLE, MANUAL_TRANSCRIPT_UNAVAILABLE, VIDEO_UNAVAILABLE, LIVE_TRANSCRIPT_UNAVAILABLE, TRANSCRIPT_REQUEST_FAILED, REQUEST_TIMEOUT, REQUEST_FAILED, INCOMPLETE_TRANSCRIPT and CHARGE_LIMIT_REACHED.

Pricing

One event is charged per complete successful transcript delivered to the Dataset. Plain text, metadata and optional SRT are included in that same event. Tracks containing only music cues return MUSIC_ONLY_CAPTIONS and are not charged as transcript events. Other error items, unavailable or missing-language results, empty or rejected incomplete transcripts, and charge-limit results are also not charged as transcript events.

Apify tier	Price per complete transcript	Equivalent per 1,000 complete transcripts
Free	$0.003000	$3.00
Bronze	$0.002700	$2.70
Silver	$0.002400	$2.40
Gold	$0.002250	$2.25
Platinum	$0.002250	$2.25
Diamond	$0.002250	$2.25

The billing unit is one delivered transcript; the 1,000-unit column is a comparison equivalent. Apify resolves the applicable account tier. A transcript charge is recorded after the complete item is written. See the Pricing tab and your account terms for any additional platform charges.

Related Actors

Actor	Best for
Social Blade Multi-Platform Analytics Scraper	Add public creator audience, views, rankings and growth metrics to YouTube content research.
PDF Text Extractor - OCR, Markdown & JSON	Add text extracted from documents to a search or RAG workflow that also uses video transcripts.

Limits and data quality

The limit is 5,000 input video URLs, not 5,000 segments per video. Processing is sequential with results saved as they finish; allow enough run time and budget for large inputs. A run stopped or timed out by the platform contains only the results already written.

“Complete” means all text-bearing segments of the selected caption response, not a guarantee that the source captions cover every spoken word. The Actor does not transcribe audio or correct recognition errors. A track made only of recognizable music cues is reported as MUSIC_ONLY_CAPTIONS with no transcript charge. When a track also has written words, its music cues, intentional repetitions and overlaps are preserved. Whitespace is normalized without summarizing the text.

Ongoing, upcoming or still-processing live streams return LIVE_TRANSCRIPT_UNAVAILABLE because their captions cannot yet represent a finished video. Finished recordings can be processed when captions are available. Private, removed, age-restricted or network-restricted videos may fail.

No successful transcript is truncated to fit a size limit. Caption responses above 16 MB or complete serialized results above the 9 MB safety budget return INCOMPLETE_TRANSCRIPT without partial text. Optional SRT increases result size; disabling it can help a large result fit while keeping all transcript segments.

Source durations are preserved when supplied. Otherwise the next segment start is used; a final unknown duration is zero. SRT is withheld if any cue has no positive usable duration; no end time is invented. SRT text escapes literal angle brackets and ampersands for subtitle rendering.

Temporary request failures receive at most two extraction attempts. When a caption response is rate-limited, the Actor may also try another format of the same caption track and language; it never silently switches languages. YouTube rate limits and caption availability can still cause errors. Review error rows and retry those videos later rather than treating a successful run status as proof that every video succeeded.

Frequently asked questions

What happens if my language is unavailable?

The item has LANGUAGE_UNAVAILABLE, with advertised languages when available. A base language may use a matching regional track, but an unrelated language is never silently substituted. YouTube-translated tracks can be used under any and are labeled isTranslated: true; this is not an independent translation service.

Can I require human-provided subtitles?

Yes. Select manual_only. If no matching manual track exists, the result has MANUAL_TRANSCRIPT_UNAVAILABLE when a matching automatic or translated alternative is advertised. If no track in the language exists, it has LANGUAGE_UNAVAILABLE.

Does this download video or generate AI summaries?

No. It retrieves available captions and metadata, not video or audio media. It does not generate summaries, embeddings or speech-to-text transcripts. Plain text and timestamps can be used in your own downstream tools.

Why is SRT null on a successful transcript?

SRT is disabled by default. If enabled but a cue lacks a usable end time, exportWarnings explains why SRT could not be produced. All transcript text and segments remain intact.

What happens when the run budget is exhausted?

Remaining videos receive CHARGE_LIMIT_REACHED without transcript content. No new caption request is made when the remaining budget is insufficient before processing starts.

Responsible use

Use public content in accordance with applicable source terms, copyright and privacy requirements. You are responsible for ensuring that retrieval and reuse of transcript text is lawful and appropriately attributed where required.

Support

For unexpected results, open an issue in the Actor's Issues tab with the run ID, a small reproducible public video URL, selected language and error code. Do not include credentials, cookies or private content.

For web usage, API examples and a sanitized sample Dataset, see the YouTube transcript scraper repository.