YouTube Subtitle Scraper — Fast & Cheap

Extract YouTube transcripts and subtitles (manual and auto-generated) together with full video metadata — without a browser. One video takes about 1 second and a fraction of a cent of compute, because every request is a plain HTTP call to YouTube's own data.

This is a YouTube transcript scraper built for scale: pipe transcripts into LLM/RAG pipelines, content repurposing tools, accessibility workflows and bulk research — with SRT/VTT export and no browser, no CAPTCHA, no wasted proxy spend.

Use Cases

• LLM & RAG pipelines — turn videos into text to summarize, quote, embed or index
• Content repurposing — video → blog post → social thread, automated on a schedule
• Bulk transcript research — transcribe a whole channel, playlist or corpus of talks for analysis
• Accessibility & subtitles — generate SRT/VTT files for videos that lack captions, in the languages you need
• Sentiment & topic analysis — analyze transcripts at scale with your own NLP models
• SEO & content mining — mine keywords and topics from competitors' videos

How It Works

Every request is a plain HTTP call — no headless Chromium, so the actor avoids the 5–15 seconds per video that browser-based subtitle scrapers pay. YouTube blocks cloud IPs, so the actor automatically uses the Apify residential proxy (a few hundred KB per video) and rotates sessions on rate limits. Extraction runs in parallel (default 10 concurrent videos, up to 20).

Subtitle track selection is graceful: it requests your chosen language, then falls back to any manual track, then any auto-generated track. Videos without subtitles (live streams, disabled captions) are reported as error items and never charged.

Extract YouTube Transcripts And Subtitles Without The Official API

YouTube's official Data API caption endpoints only work for videos you own — they cannot fetch captions for third-party videos, and they carry quotas. This actor reads the same transcript data the YouTube page uses, at scale, with no OAuth and no quota management on your side.

• Input flexibility — watch, shorts, youtu.be, /embed/, /live/ URLs and even bare 11-character video IDs
• Playlists — switch on playlist expansion to process every video in a playlist (with a safety cap)
• Formats — plain transcript text, timestamped transcript, and ready-to-use SRT and VTT subtitle files
• Metadata — title, channel, views, length, publish date and category, alongside each transcript

Input

Field	Type	Required	Default	Notes
startUrls	array	Yes	one sample video	Video URLs (watch, shorts, youtu.be) — or playlist URLs when Expand playlists is on
language	string	No	en	Transcript language. Falls back to any manual track, then auto-generated. any = best available
timestamps	boolean	No	false	Prefix each transcript line with [h:mm:ss] and include per-segment data
includeSrt	boolean	No	false	Add a ready-to-use SRT subtitle file as a text field
includeVtt	boolean	No	false	Add a ready-to-use WebVTT subtitle file as a text field
expandPlaylists	boolean	No	false	Treat playlist URLs as a list of videos and process every video
maxPlaylistVideos	integer	No	100	Safety cap when expanding playlists
maxVideos	integer	No	0	Hard cap on total videos in a run. 0 = no cap
concurrency	integer	No	10	Videos extracted in parallel (1–20)
proxyConfiguration	object	No	Apify residential proxy	Defaults to residential because YouTube blocks cloud IPs

Example Input

{
  "startUrls": [
    { "url": "https://www.youtube.com/watch?v=SieELGaJV8k" },
    { "url": "https://youtu.be/dQw4w9WgXcQ" }
  ],
  "language": "en",
  "timestamps": false,
  "includeSrt": false,
  "concurrency": 10
}

Example Output

Field	Description
url, video_id	Canonical video URL and ID
title, channel_name, channel_id	Video and channel identity
views, length_seconds, publish_date, category	Raw metadata as reported by YouTube
transcript_language, transcript_language_code, is_auto_generated	Which subtitle track was used
segment_count, transcript_chars	Size of the transcript
transcript	Full transcript text (with [h:mm:ss] timestamps when enabled)
segments	Per-line {start, duration, text} — only when Timestamped transcript is on
srt, vtt	Ready-to-use subtitle files — optional
error	Failure reason — only for videos that could not be processed

{
  "url": "https://www.youtube.com/watch?v=SieELGaJV8k",
  "video_id": "SieELGaJV8k",
  "title": "Example video",
  "channel_name": "Example Channel",
  "views": 123456,
  "length_seconds": 612,
  "transcript_language": "English",
  "is_auto_generated": false,
  "segment_count": 210,
  "transcript_chars": 8421,
  "transcript": "Welcome back to the channel..."
}

Pricing

Pay-per-event, billed per successful video — failed videos are never charged:

Event	Price	When
result	$0.01	One extracted transcript with metadata
long-video	$0.02	Add-on, charged for videos longer than 20 minutes
apify-actor-start	$0.00005	Per GB of memory when the run starts (512 MB = $0.00005)

Example: a 10-video batch of regular videos costs $0.10; a single 3-hour podcast costs $0.03. For comparison, a 10-minute video with a browser-based YouTube scraper on the Store commonly costs several times more in proxy + compute.

Integrations And Automation

• Apify API — call the actor from your app and read the dataset.
• Webhooks — kick off downstream summarization or indexing when a run finishes.
• Scheduled runs — monitor channels/playlists daily or weekly to capture new uploads automatically.
• Zapier / Make / n8n — feed new transcripts into Notion, Google Sheets, Airtable or a vector store.
• Export — JSON, CSV, Excel; transcript text stays clean and ready for LLMs.

Related Actors

• YouTube Scraper — the largest YouTube actor for channel, video and comment metadata; pair with this one for transcripts
• YouTube Comments Scraper — audience sentiment to complement your transcript analysis
• TikTok Scraper — short-form video data for cross-platform content research
• Instagram Scraper — multi-platform content dataset building
• Website Content Crawler — extend your RAG corpus from video transcripts to full web pages

FAQ

Does YouTube have an official transcript API?

YouTube's official Data API caption endpoints only work for videos you own, so they cannot fetch captions for third-party videos. For everyone else, the options are brittle scraping of YouTube's internal endpoints or a hosted service — this actor does that work reliably, at scale, charging only for successful videos.

Why is this actor faster and cheaper than other YouTube transcript scrapers?

It uses plain HTTP, not a headless browser. Browser-based subtitle scrapers pay for proxy traffic and burn 5–15 seconds per video; this actor does not, so it is faster and dramatically cheaper per video.

Which formats can I get?

Plain transcript text, timestamped transcript, and ready-to-use SRT and VTT subtitle files — all from the same run.

Can I transcribe a whole playlist or channel?

Yes. Turn on Expand playlists to process every video in a playlist (with maxPlaylistVideos as a safety cap); feed a list of video IDs to process a channel's videos in bulk.

What happens to videos with no captions?

They are reported as error items and are never charged. Live streams and videos with disabled captions typically have no subtitle track available.

Does auto-generated text work?

Yes. Auto-generated subtitles exist for most videos, so coverage is very high. The track used is reported in is_auto_generated.

How do I use this for an LLM / RAG pipeline?

Set language to your language and read the transcript field (or segments for timestamped data), then chunk and embed it in your vector store. Schedule the actor to keep the corpus fresh.

Can I export transcripts to CSV or JSON?

Yes — JSON, CSV, Excel and more from the dataset tab or the API.

Legal & Compliance Disclaimer

This actor is an independent tool and is not affiliated with, endorsed by, or sponsored by YouTube or Google. It accesses only publicly available subtitle/transcript data and video metadata; it does not bypass login and does not solve CAPTCHAs.

You are responsible for complying with YouTube's Terms of Service and all applicable laws, including copyright rules — transcripts are derived from creators' content, so respect the rights of video owners when you store, republish or train on them.

For AI Agents & LLM Apps

Purpose: returns YouTube transcripts/subtitles plus video metadata as structured JSON, one item per successfully processed video, for RAG, summarization and bulk content analysis.

Minimal working input:

{ "startUrls": [{ "url": "https://www.youtube.com/watch?v=SieELGaJV8k" }] }

Variant — timestamped transcript with subtitle files:

{
  "startUrls": [{ "url": "https://www.youtube.com/watch?v=SieELGaJV8k" }],
  "language": "en",
  "timestamps": true,
  "includeSrt": true,
  "includeVtt": true
}

Variant — expand a playlist:

{
  "startUrls": [{ "url": "https://www.youtube.com/playlist?list=PLxxxxxxxx" }],
  "expandPlaylists": true,
  "maxPlaylistVideos": 50,
  "concurrency": 10
}

Output fields: url, video_id, title, channel_name, channel_id, views, length_seconds, publish_date, category, transcript_language, transcript_language_code, is_auto_generated, segment_count, transcript_chars, transcript, segments, srt, vtt, error.

Behaviors an agent should know:

• startUrls is required and accepts watch/shorts/youtu.be/embed/live URLs or bare 11-char video IDs.
• language defaults to en; if unavailable it falls back to any manual track, then auto-generated. Use any for best-available.
• segments is only populated when timestamps is true; srt only when includeSrt is true; vtt only when includeVtt is true.
• Playlist URLs are only expanded when expandPlaylists is true (otherwise treated as a single target); maxPlaylistVideos caps the expansion.
• maxVideos is a hard total cap; 0 means no cap. Default concurrency is 10 (max 20).
• Videos without captions return an error item and are never billed.
• Billing is pay-per-event: result $0.01 per successful transcript, long-video +$0.02 for videos over 20 minutes, plus a tiny per-GB start charge.

SEO Keywords

youtube transcript scraper, youtube transcript api, youtube subtitles scraper, youtube captions downloader, extract youtube transcript, bulk youtube transcript, youtube transcript bulk download, youtube to text, youtube srt, youtube vtt, youtube subtitle downloader, youtube transcript for llm, youtube transcript for rag, transcript api alternative, cheap youtube transcript api, youtube transcript without api key, youtube channel transcript, youtube playlist transcript, auto generated captions scraper, transcript extraction, youtube subtitle generator, no browser youtube scraper, batch youtube transcripts, youtube content research, video to text