YouTube Transcript Scraper (bulk, fast & reliable)

Fetch transcripts for up to 500 YouTube videos in one run. Built for AI agents, content pipelines and bulk analysis: strict output schema, timestamped snippets, optional SRT export, language selection, and honest billing — you only pay for transcripts that are actually delivered.

Why this scraper

• Only successful results are charged. Videos without captions, deleted videos or temporary blocks show up as clear error records at no cost.
• Built for bulk. Paste up to 500 URLs or video IDs per run; every result carries the video ID so you can join it back to your own data.
• Agent-ready output. A stable, machine-readable schema (no HTML, no surprises) with explicit error codes — ideal as a tool behind an LLM or automation flow.
• Language control. Preference order (e.g. ["nl", "en"]), and manual captions are preferred over auto-generated ones whenever both exist.
• Lightweight route. This actor uses a slim API path instead of loading the full video page, which keeps it fast and keeps the price low.

Who this is for

Teams building AI tools on video content: RAG pipelines, summarisers, research assistants. Content teams repurposing video into articles and newsletters. Anyone who needs transcripts for hundreds of videos and cannot babysit a scraper.

If you need one transcript, once, a browser extension is cheaper. This actor earns its place from roughly the tenth video onwards, and from the moment reliability matters.

Who this is not for

Read this before you buy. The actor returns the captions YouTube has for a video, manual or auto-generated, in the languages YouTube offers. It does not transcribe audio itself, so a video without any caption track comes back as a free NO_TRANSCRIPT record, and it does not translate: languages picks among the tracks that exist. It also does not download the video or audio. For speech-to-text on caption-less videos you need a transcription service, not a scraper.

Input example

{
    "videos": [
        "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
        "https://youtu.be/JGwWNGJdvx8",
        "kJQP7kiw5Fk"
    ],
    "languages": ["en"],
    "preferManualCaptions": true,
    "includeTimestampedSnippets": true,
    "includeSrt": false
}

Accepted video formats: full watch URLs, short links (youtu.be), shorts, embeds, or bare 11-character video IDs.

Output example

{
    "videoId": "dQw4w9WgXcQ",
    "status": "ok",
    "title": "Rick Astley - Never Gonna Give You Up",
    "channel": "Rick Astley",
    "durationSeconds": 213,
    "language": "en",
    "languageName": "English",
    "isAutoGenerated": false,
    "availableLanguages": ["de", "en", "es", "fr", "ja", "pt"],
    "text": "We're no strangers to love ...",
    "snippets": [
        {"text": "We're no strangers to love", "start": 18.88, "duration": 3.96}
    ]
}

A failed video produces an error record instead (and is not charged):

{
    "videoId": "9bZkp7q19f0",
    "status": "error",
    "errorCode": "NO_TRANSCRIPT",
    "error": "Video 9bZkp7q19f0 has no caption tracks"
}

Error codes: INVALID_VIDEO_ID, VIDEO_UNAVAILABLE, NO_TRANSCRIPT, LANGUAGE_NOT_FOUND, BLOCKED. INPUT_TRUNCATED appears once when your input is longer than this actor accepts.

Related actors from steadydata

• youtube-video-details: the metadata for the same videos: views, tags, category
• youtube-channel-videos: every video of a channel, to feed this actor a list
• youtube-comments: what viewers said, without author names

Pricing

Pay per event: one transcript-fetched event per delivered transcript. No charge for failed videos, no charge for videos without captions, no separate platform-usage surcharge.

FAQ

Does it work with auto-generated captions? Yes. Set preferManualCaptions to false to take auto-generated tracks even when a manual track exists; by default manual tracks win.

Can I get subtitles as a file? Set includeSrt to true and every result carries a ready-to-use srt field.

What about very long videos? Duration does not change the price: one delivered transcript is one event.

Is personal data collected? No. The actor returns public caption text and video metadata only.

What happens when YouTube changes something? Transcript sources break from time to time; that is the nature of this work. The difference here is response: the actor is monitored daily and fixed fast, and while it is broken you are not charged, because only delivered transcripts cost anything.

What if a video has no captions at all? You get an explicit NO_TRANSCRIPT record for that video and pay nothing for it. Your run continues with the rest of the list.