YouTube Transcript Scraper - subtitles & captions for AI / RAG

Extract YouTube transcripts, subtitles and captions in bulk - from single videos, whole channels or playlists - as clean JSON ready for LLMs, RAG pipelines, search and analysis. No login, no API key, no cookies.

Built for reliability: automatic retries with backoff, per-video error isolation (one failing video never breaks the run), automatic datacenter-to-residential proxy fallback if YouTube blocks the IP, and a clear error code for every video that has no subtitles.

What you get

For every video:

Field	Description
videoId, url	Video identifiers
title, channel, channelUrl, durationSeconds, publishedAt, viewCount	Video metadata
languageCode, language	Language of the transcript returned
type	manual (uploaded subtitles), auto (YouTube speech recognition) or translated
text	Full transcript as one string
segments	Timestamped segments {start, duration, text}
chapters	Video chapters (with their transcript text in clean/rag mode)
chunks	In rag mode: ~1,500-character chunks with start times, ready for embeddings
status, errorCode, error	ok, no_captions or error with a readable reason

Features

• Smart language fallback: manual subtitles in your language → auto-generated in your language → YouTube machine translation → original language of the video.
• Videos, channels, playlists: paste URLs (watch, youtu.be, shorts, embed, live) or plain IDs; add a channel/@handle/playlist to get its latest N videos.
• Three output formats: full (text + segments), clean (readable text without [Music] / >> noise - ideal for LLM prompts), rag (clean text + timestamped chunks for vector databases).
• Chapters with per-chapter transcript text for structured summaries.
• Reliable: retries with exponential backoff, concurrency control, automatic Apify Proxy fallback on IP blocks.
• Pay only for results (pay-per-event pricing planned): videos without subtitles are free.

Input example

{
  "videoUrls": ["https://www.youtube.com/watch?v=8jPQjjsBbIc", "M7lc1UVf-VE"],
  "channelOrPlaylistUrls": ["https://www.youtube.com/@TED"],
  "maxVideosPerSource": 5,
  "languages": ["en", "it"],
  "outputFormat": "rag"
}

Output example

{
  "videoId": "8jPQjjsBbIc",
  "title": "How to stay calm when you know you'll be stressed | Daniel Levitin | TED",
  "channel": "TED",
  "durationSeconds": 740,
  "languageCode": "en",
  "type": "manual",
  "status": "ok",
  "text": "A few years ago, I broke into my own house. ...",
  "segments": [{"start": 13.24, "duration": 2.56, "text": "A few years ago,\nI broke into my own house."}],
  "chapters": [{"title": "Introduction", "start": 0, "end": 85}]
}

Use cases

• Feed YouTube video content to ChatGPT / Claude / LLM agents and RAG knowledge bases
• Summarise lectures, podcasts, interviews and webinars at scale
• SEO and content research, topic and keyword analysis on competitor channels
• Dataset building for NLP and speech research, accessibility and archiving
• Repurpose videos into blog posts, newsletters and social posts

Reliability and proxies

YouTube blocks most datacenter IPs, so the connection method decides how many videos succeed. How the Actor escalates, per video:

1. Datacenter Apify Proxy (default). Included in the free Apify plan. In our cloud tests about 75% of videos that have subtitles succeeded (3 of 4, small sample); direct connections without a proxy failed every time (0 of 3).
2. RESIDENTIAL proxy (automatic fallback, useResidentialFallback, on by default). If the datacenter proxy keeps getting blocked for a video, the Actor retries that video through the Apify RESIDENTIAL group. Residential proxies are billed to your Apify account by traffic. A transcript is roughly 30-60 KB, so we estimate about 0.0005 USD per video at ~8 USD/GB (estimate from typical transcript size and public proxy pricing, not a guarantee). Residential IPs are recommended if you need more than 95% success; we have not yet measured the exact rate on large batches.
3. If RESIDENTIAL is not available on your account, the run does not crash: the video is saved with status: error and errorCode: BLOCKED. To fix it, enable residential proxies in your Apify plan, or select the RESIDENTIAL group in the Proxy configuration input to use it from the start. You can also simply re-run the failed videos later.

Set useResidentialFallback to false to never use residential traffic. Videos without any subtitles are reported as NO_CAPTIONS and are not retried.

Limitations (honest notes)

• Only videos that have subtitles (manual or auto-generated) can be transcribed. Videos with captions disabled return status: no_captions and cost nothing.
• Age-restricted, private and members-only videos need a login and are reported as errors - the Actor never logs in.
• Machine translation is rate-limited by YouTube and may fall back to the original language.
• YouTube may block datacenter IPs, so some videos can end as BLOCKED - see Reliability and proxies above.

FAQ

Do I need a YouTube API key? No. Does it work for auto-generated captions? Yes (type: auto). Can I get a transcript in another language? Set languages; if the video has no such track the Actor tries YouTube's translation, then falls back to the original language. Is it legal? It reads publicly available captions; you are responsible for complying with YouTube's terms and applicable law for your use case.

Use via API, schedules, webhooks, or integrations (Make, n8n, Zapier, LangChain, LlamaIndex) like any Apify Actor.

Who is this for?

AI developers building RAG or summarisation tools, researchers and students, content marketers repurposing videos, and SEO teams analysing competitor channels.

More free tools by tomstudio

• ATS Jobs Scraper (Workday, Greenhouse, Lever, Ashby...)
• TED Tenders Monitor (EU & Swiss public procurement)
• SEC EDGAR Filings Monitor
• Domain Availability & RDAP Bulk Checker
• App Store Reviews Scraper
• Hacker News Search & Comments
• Wayback Machine Snapshots
• RSS & Atom Feed Reader
• YouTube Transcript Scraper

Ready-to-use examples

Copy any of these inputs into the Input tab (the same examples are saved as public Tasks on this Actor's page).

YouTube transcript: Rick Astley - Never Gonna Give You Up

Get the full timestamped transcript of a famous YouTube video.

{
  "videoUrls": [
    "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
  ],
  "languages": [
    "en"
  ],
  "outputFormat": "full"
}

YouTube transcript: Steve Jobs Stanford speech (RAG chunks)

Transcript of the Steve Jobs 2005 Stanford commencement speech, chunked for AI/RAG.

{
  "videoUrls": [
    "https://www.youtube.com/watch?v=UF8uR6Z6KLc"
  ],
  "languages": [
    "en"
  ],
  "outputFormat": "rag",
  "chunkSize": 1500
}