find_youtube_outliers
Surface high-performing videos for up to five keywords. Identify packaging patterns, overperforming thumbnails, and underexploited angles.
Remote MCP Server
YouTube MCP Server for AI Agents.
Connect Claude Code, Cursor, Windsurf, and VS Code agents to real YouTube data through a remote MCP server. Pull outliers, niche explorer channels, deep research audits, and video context without leaving your editor.
MCP Tool Call Example
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "find_youtube_outliers",
"arguments": {
"keywords": ["remote work vlogs"]
}
}
}What is MCP and why it matters.
Model Context Protocol is an open standard that turns AI agents into real operators. Instead of asking an agent to write code that calls an API, you give it a list of tools it can invoke directly. The agent decides what to call, passes arguments, reads the response, and continues reasoning.
For creators and researchers, this means you can ask Claude Code to find high-performing history niche videos, then immediately ask it to cluster the packaging patterns and propose five original scripts. The data and the reasoning happen in one place.
Autonolab MCP at a glance
- Remote server hosted by Autonolab. Nothing to install or maintain.
- Works with Claude Code, Cursor, Windsurf, and any MCP client.
- Five YouTube research tools mapped to production REST endpoints.
- Free tier: 20 calls per day, 5 per minute.
- API key authentication. No OAuth dance required.
Tools available on the server.
Each MCP tool maps to a production Autonolab endpoint. Your agent discovers them automatically and decides which to call based on your prompt.
search_niche_channels
Query Niche Explorer enriched channel data. Find small channels punching above their weight with clear audience signals.
audit_youtube_channel
Run the Channel Audit deep research AI agent against any public channel and get strategy, content gaps, and positioning notes.
find_youtube_niche
Run the Niche Finder deep research AI agent from your interests, constraints, and content goals to discover viable channel ideas.
get_channel_videos
Fetch recent uploads from any channel for repackaging research, competitor comparison, and content calendar planning.
How to connect in three steps.
No local setup, no Docker containers, no server management. Add the remote MCP endpoint to your agent client and start researching YouTube data in minutes.
01
Create an API key
Sign up for Autonolab and generate a developer API key from your dashboard. Free tier starts with 20 calls per day.
02
Add the remote MCP server
In Claude Code, Cursor, Windsurf, or any MCP-compatible client, add the Autonolab remote MCP server URL and paste your API key.
03
Ask your agent to research
Prompt your agent naturally. It will call the right MCP tool, fetch YouTube data, and continue the conversation with real context.
Code example: direct REST fallback.
If your client does not support MCP yet, you can call the same endpoints directly. Here is an example using curl.
curl
curl -X POST https://api.autonolab.com/api/v1/youtube/outliers \
-H "X-API-Key: ak_live_..." \
-H "Content-Type: application/json" \
-d '{
"keywords": ["productivity systems"],
"filters": {
"videoType": "long"
}
}'Who the MCP server is for.
If you already live inside an AI agent workflow, adding YouTube research should feel like plugging in a new sense.
Software engineers with YouTube side projects
You already use Claude Code or Cursor daily. Add a research layer without switching tabs or copying links manually.
Agentic workflow builders
Build internal research bots and automation pipelines that pull creator intelligence into Slack, Notion, or Airtable.
Data teams and creator ops
Enrich dashboards with outlier signals, channel audits, and niche opportunity scores instead of manual YouTube browsing.
Independent creators who want speed
Stop opening fifty tabs to research a video idea. Ask your agent to find proof points while you stay in the editor.
Frequently asked questions.
What is an MCP server?
MCP stands for Model Context Protocol. It is an open standard that lets AI agents discover and call external tools dynamically. Autonolab exposes YouTube research functions as MCP tools so Claude Code, Cursor, Windsurf, and other clients can fetch creator data without you writing HTTP requests by hand.
Do I need to host anything locally?
No. Autonolab runs the MCP server remotely. You only need to add the server URL to your MCP client and provide your API key. The client handles discovery, authentication, and tool execution automatically.
Which AI agents support remote MCP servers?
Claude Code, Cursor, Windsurf, and several VS Code agent extensions support remote MCP servers today. The ecosystem is growing fast. Autonolab is designed to work with any client that implements the MCP specification.
How is this different from calling a REST API?
With a REST API, you craft requests, parse JSON, and build prompts yourself. With MCP, your agent sees available tools, decides which to call, passes arguments, and consumes results within the same conversation. It removes the glue work.
What tools are available on the server?
The server currently exposes five tools: find_youtube_outliers, search_niche_channels, audit_youtube_channel, find_youtube_niche, and get_channel_videos. Each tool maps to a specific Autonolab research endpoint.
Are there rate limits?
Free accounts receive 20 calls per day and 5 per minute. Paid plans increase both limits. Rate limits protect service stability and keep the free tier usable for side projects and light research.
Can I use this in automated scripts or CI pipelines?
Yes. While MCP shines in conversational clients, you can also call the underlying REST endpoints directly with your API key. The MCP layer is just the most convenient interface for agentic workflows.
What data does the server access?
The server queries publicly available YouTube metadata, rankings, and Autonolab proprietary signals such as outlier scores, niche opportunity ratings, and deep research summaries. It does not access private video data or authenticated YouTube accounts.
Related developer pages
Plug your agent into real YouTube data.
Create a free API key, add the Autonolab MCP server to your client, and start researching outliers, niches, and audits without leaving your editor.