YouTube data for AI agents
Video detail, channel feeds, comments, subtitles, search and trending, plus public channel contact lookup. One key, and no daily quota units to budget.
Copy the setup prompt.
Your agent does the rest.
Fetch and execute the appropriate instructions to set me up for superagnt_ from https://mcp.superagnt.com/agent-setup/prompt.md» once it is connected, run this so your agent can switch the source on itself.
agnt_tools_enable({ families: ['data:youtube'] })Everything your agent can pull
Title, description, view and like counts, publish date, duration, channel and thumbnails for a video id.
Caption tracks for a video, so an agent can read what was actually said instead of inferring from the title.
Comment threads with continuation paging, for reading audience reaction at volume.
Channel details, uploads, shorts and in-channel search, with continuation tokens for deep history.
Keyword search with language, country and ordering, plus trending by country and section.
Audio track detail and every video using a given track.
Public business email for a channel by URL or channel id, for outreach workflows.
tool names08as your agent sees them
data_youtube_video_detailsdata_youtube_video_subtitlesdata_youtube_video_commentsdata_youtube_channel_detailsdata_youtube_channel_videosdata_youtube_youtube_searchdata_youtube_trending_videosdata_youtube_get_channel_email_by_urlThree jobs it can start on today
A research agent that actually watches
The agent searches a topic, pulls subtitles for the top results, and summarises what was claimed with timestamps and a link back to each source video.
Creator outreach list
From a keyword search the agent collects channels above a subscriber floor, resolves each public business email, and writes a qualified list into your CRM.
Launch reaction digest
After a product video goes live, a schedule reads new comments every hour, clusters them by theme, and posts the top questions and complaints to Slack.
What comes back
Your agent reads this response directly, with no parsing layer of your own in between. The same JSON comes back over plain HTTP, so a cron job or a script sees exactly what the agent sees.
https://mcp.superagnt.com/mcptool calldata_youtube_video_details{
"video_id": "kR3xB2QpV1s"
}{
"video_id": "kR3xB2QpV1s",
"title": "Building an inbox agent in 20 minutes",
"channel_id": "UC0000000000000000000A",
"channel_title": "North Loop Labs",
"published_at": "2026-07-14T15:02:11Z",
"length_seconds": 1247,
"view_count": 84102,
"like_count": 3110,
"...": "description, keywords, thumbnails, is_live"
}What it costs to run
Calls in YouTube are priced individually and deducted from your credit balance, and only a successful call is billed. The per-endpoint figures live in the reference docs, and every billable response reports its own cost in meta.costCents.
One balance covers this source and every other one, so there is no per-source subscription and no upstream account of your own to keep alive.
how credits work →Before you wire it up
01Do I need a YouTube Data API key or quota?
No. There are no daily quota units to plan around and no cloud project to manage. You pay per call in credits.
02How does my agent get these tools?
Connect to https://mcp.superagnt.com/mcp and call agnt_tools_enable with data:youtube.
03Can it read what was said in a video?
Yes. data_youtube_video_subtitles returns the caption track, which is what makes summarising and quoting a video reliable.
04How is usage billed?
Calls consume platform credits based on provider and endpoint. See credit costs in the dashboard and on each endpoint page.
05What about rate limits?
Rate limits depend on your plan. The API returns standard HTTP status codes; retry with backoff on 429 responses.
Point your client at https://mcp.superagnt.com/mcp and your agent has this, plus everything else on one balance.