YouTube Transcripts & Metadata Scraper

YouTube transcript API, captions and subtitles — extract clean, timestamped transcripts and metadata from YouTube videos, Shorts, channels and playlists, reliably and at scale. Output as text, markdown, SRT or VTT, ready for RAG, LLM and AI agent pipelines.

What does YouTube Transcript Scraper do?

It turns any public YouTube video into structured text: the full transcript segmented by timestamp, plus video metadata. It works on single videos, full URL batches, Shorts, entire playlists and the latest videos of a channel, in the languages you choose — and it can translate captions when a language is missing.

It does not transcribe audio: it extracts captions/subtitles that already exist on the video (auto-generated or uploaded). It's a reliable YouTube transcript API alternative that needs no Google API key and no OAuth.

Why scrape YouTube transcripts?

Video is where a huge amount of knowledge lives, but it isn't searchable, embeddable or queryable. Transcripts make it all three.

• Feed RAG and LLM pipelines. Drop transcripts straight into a vector store, a LangChain or LlamaIndex loader, or an agent's context window. Each row is AI-ready with segments, fullText and markdown.
• Repurpose video into text. Turn webinars, interviews and tutorials into blog posts, newsletters, subtitles and clip scripts.
• Research and analyse at scale. Pull hundreds of videos on a topic and run topic classification, entity extraction, sentiment or search over the text.
• Publish transcripts for SEO and accessibility. Make video content indexable and provide a text alternative for viewers who need it.
• Monitor channels and playlists. Run on a schedule to keep a keyword or knowledge base fresh.

Built for reliability and scale

• Batch thousands of URLs in one run, with configurable concurrency.
• Per-item status/error — one bad video never breaks the run.
• Retries with exponential backoff and Apify Proxy support to avoid IP blocks.
• Multiple output formats: segments, text, markdown, SRT, VTT.

What data can it extract?

Every video produces one dataset row:

Field	Type	Description
videoId	string	YouTube video ID.
title	string	Video title.
channelName / channelId	string	Publishing channel.
publishDate	string	Publication date.
durationSec	number	Duration in seconds.
viewCount	number	View count at run time.
tags	array	Video keywords/tags.
description	string	Video description.
thumbnailUrl	string	Max-resolution thumbnail.
language	string	Transcript language (ISO 639-1).
isGenerated	boolean	Whether captions are auto-generated.
translatedFrom	string	Source language if translated.
segments	array	{start, end, text} timestamped segments.
fullText	string	Full transcript as one string.
markdown	string	Transcript as markdown with the video link.
srt / vtt	string	Subtitle formats (when requested).
status / error	string	Per-item outcome.

How to scrape YouTube transcripts (step-by-step)

1. Open the Actor in Apify Console or call it via API.
2. Add Video URLs (watch/youtu.be/Shorts), or add Channel URLs / Playlist URLs to expand them automatically.
3. Set Language priority (e.g. ["en", "es"]).
4. Choose the Output format (segments, text, markdown, srt, vtt).
5. Start the run. Watch the log and results in the Output tab.
6. Download the dataset as JSON, CSV or Excel, or read it via the API. Schedule the run to keep your data fresh.

How much does it cost to scrape YouTube?

You pay per event, not per compute time. Prices:

• $3 per 1,000 transcripts (transcript-extracted)
• $0.50 per 1,000 videos of metadata (metadata-enriched)

Failed videos and videos without captions are not charged. On Apify's free plan you get monthly credits to test, and this Actor caps free-plan runs to the first few videos, so you can validate the output before paying.

Example: 1,000 videos with transcripts and metadata ≈ $3.50.

Input

See the Input tab for full configuration. Key fields:

Field	Type	Default	Description
videoUrls	array	—	Video/Shorts URLs (watch, youtu.be, shorts).
channelUrls	array	—	Channel URLs, expanded to their latest videos.
playlistUrls	array	—	Playlist URLs, expanded to their videos.
languages	array	["en"]	ISO 639-1 codes in priority order.
includeAutoGenerated	boolean	true	Fall back to auto-generated captions.
outputFormat	string	segments	segments, text, markdown, srt, vtt.
includeMetadata	boolean	true	Collect video metadata.
maxItems	integer	0	Max videos. 0 = unlimited (free plan capped).
maxRetries	integer	3	Per-request retries.
concurrency	integer	5	Parallel video workers.
proxyConfiguration	object	Apify Proxy	Apify Proxy (auto) by default; set Residential at high volume if blocked.

Output

One JSON object per video. Example:

{
  "videoId": "dQw4w9WgXcQ",
  "title": "Rick Astley - Never Gonna Give You Up (Official Video)",
  "channelName": "Rick Astley",
  "publishDate": "2009-10-25",
  "durationSec": 213,
  "viewCount": 1820610046,
  "language": "en",
  "isGenerated": false,
  "segments": [{"start": 1.36, "end": 3.04, "text": "We're no strangers to love"}],
  "fullText": "We're no strangers to love ...",
  "status": "ok",
  "error": null
}

Integrate (API / SDK / MCP)

from apify_client import ApifyClient

client = ApifyClient("<APIFY_TOKEN>")
run = client.actor("youtube-transcript-scraper").call(run_input={
    "videoUrls": ["https://www.youtube.com/watch?v=dQw4w9WgXcQ"],
    "languages": ["en", "es"],
    "outputFormat": "srt",
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["videoId"], item["status"])

This Actor is callable as a tool for AI agents through the Apify MCP server (https://mcp.apify.com), so an agent can fetch a transcript on demand — no glue code.

FAQ

Can I download subtitles as SRT or VTT?

Yes. Set outputFormat to srt or vtt and each row includes subtitles ready to use.

Does it work with YouTube Shorts and playlists?

Yes. Shorts URLs are accepted and playlist/channel URLs are expanded into their videos automatically.

Can I get transcripts in another language?

Yes. Pass a language priority list; if the language isn't available, the transcript is translated when possible (the translatedFrom field tells you the source language).

Does it transcribe videos without captions?

No. It extracts existing captions/subtitles; it does not perform speech-to-text.

How many videos can I process per run?

There is no fixed limit — batch as many URLs as you need, subject to maxItems and your plan.

Disclaimers & support

Use this Actor only for lawful purposes and content you are permitted to process. Transcripts may contain copyrighted material or personal data; keep source attribution and comply with YouTube's terms and your local laws. For bug reports or feature requests, open an issue on the Actor's page.