Platform · YouTube
YouTube Data API for AI Agents
One API and one MCP server for every YouTube read operation your product needs. No quota units, no OAuth dance, no approval process.
The official YouTube Data API v3 now separates common operations into different daily limits: new projects receive 100 search.list calls per day alongside a 10,000-unit general quota pool. Its caption workflow requires OAuth, and captions.list returns metadata rather than the transcript text. CreatorCrawl gives you channels, videos, comments, transcripts, playlists, search, and trending Shorts under pay-as-you-go credits that never expire. Every endpoint is exposed as a native MCP tool so your Claude, Cursor, or Zed agent can read YouTube data in natural language.
Free YouTube tools
Try CreatorCrawl in your browser
Test live public YouTube data one URL at a time. No login or API key required.
Why pick CreatorCrawl over the YouTube Data API v3?
The official YouTube Data API v3 is the first place most developers look. Here is why teams building real products switch off it.
YouTube Data API v3
- New projects receive 100 search.list calls per day
- A separate 10,000-unit daily pool applies to general operations
- Caption metadata and downloads require OAuth, and captions.list does not return transcript text
- Commercial review is required for higher quotas
- OAuth overhead for read-only public data
CreatorCrawl
- No approval process, sign up in 60 seconds
- Pay-as-you-go credits, never expire, no subscription
- Native MCP server for Claude, Cursor, Zed
- Same API covers 5 other platforms under one key
- Real-time data on every request, 50 free credits on signup
YouTube endpoints
10 endpoints covering profiles, content, engagement, and search. Every endpoint is exposed as an MCP tool so your AI agent can call it in natural language.
/youtube/channelChannel
Fetch a YouTube channel, subscriber count, total views, country, join date, and description.
/youtube/channel/videosChannel Videos
List long-form videos from any channel with pagination.
/youtube/channel/shortsChannel Shorts
Dedicated endpoint for a channel's YouTube Shorts.
/youtube/videoVideo
Title, description, views, likes, comment count, thumbnail, publish date, and duration for a single video.
/youtube/video/commentsVideo Comments
Paginated comments with replies, like counts, and author info. No quota accounting.
/youtube/video/transcriptVideo Transcript
Plain-text transcript plus timestamped segments without the official captions OAuth and download workflow.
/youtube/playlistPlaylist
List videos inside any public playlist with pagination.
/youtube/searchSearch
Keyword search across the whole platform with filters.
/youtube/search/hashtagSearch by Hashtag
Find videos tagged with a specific hashtag.
/youtube/trending/shortsTrending Shorts
Currently trending Shorts by region. Useful for content research and trend detection.
Connect YouTube to your AI agent in 60 seconds
The CreatorCrawl MCP server exposes all 10+ YouTube endpoints as native tools. Add the server URL to Codex, Claude, Cursor, or another OAuth-capable MCP client and your agent can read YouTube in natural language.
{
"mcpServers": {
"creatorcrawl": {
"url": "https://app.creatorcrawl.com/api/mcp"
}
}
}Save the config and connect. Your client opens CreatorCrawl in your browser for sign-in and authorization, then refreshes OAuth tokens automatically. Ask your agent something like "Pull the last 20 posts from a YouTube account and summarise the themes."
What teams build with YouTube data
Build an LLM video summariser
Pull the transcript endpoint for any YouTube URL and feed it straight into Claude or GPT. The plain-text field means zero parsing. Works with Shorts and long-form video.
Competitor channel monitoring
Track rival channels for new uploads, view count changes, and comment sentiment. Paginate channel videos on a schedule and diff against the previous run.
Niche research for content planning
Search the platform for your niche, rank videos by engagement, and extract the transcripts of the top 10. Feed the text into an LLM to surface repeatable hook patterns.
AI agents that reason about YouTube
Point Claude or Cursor at the CreatorCrawl MCP server and your agent can search, fetch, and analyse YouTube content in natural language without writing any API client code.
Frequently asked questions about the YouTube API
How is this different from the official YouTube Data API v3?
CreatorCrawl uses one API key and returns transcript text directly. New Google Cloud projects receive 100 search.list calls per day plus a separate 10,000-unit general pool, while the official caption list and download workflow requires OAuth. One CreatorCrawl API call is one credit and credits never expire.
Do you support YouTube Shorts?
Yes. There is a dedicated Channel Shorts endpoint for listing a creator's Shorts, and a Trending Shorts endpoint for the country-level trending feed. The Video and Video Comments endpoints work with any video ID, Shorts included.
Can I extract YouTube transcripts?
Yes. The Video Transcript endpoint returns timestamped segments and a plain-text field. The official YouTube captions resource can list and download caption tracks for authorised workflows, but captions.list does not include the caption text and the workflow requires OAuth.
Is there a free tier?
50 credits free on signup, no credit card required. Each API call costs 1 credit. After the free pack, pay-as-you-go credit bundles start at $29 for 5,000 credits. Credits never expire.
Does it work with AI agents like Claude and Cursor?
Yes. CreatorCrawl ships a native MCP server that exposes every YouTube endpoint as a tool. Drop one config block into Claude Desktop, Claude Code, Cursor, or Zed and your agent can call YouTube tools in natural language without any glue code.
Is it legal to read YouTube data this way?
CreatorCrawl only accesses publicly available YouTube data, the same data any logged-out browser can see. We do not bypass private content, login walls, or paid features. You are responsible for complying with YouTube Terms of Service and applicable copyright rules for your downstream use.