Yutori MCP

MCP tools and skills for web monitoring, deep research, and browser automation — powered by Yutori's web agentic tech.

You can use it with Claude Code, Codex, Cursor, VS Code, ChatGPT, OpenClaw, and other MCP hosts.

Features

Capabilities:

• Scouting — Monitor the web continuously for anything you care about at a desired frequency
• Research — Run one-time deep web research tasks
• Browsing — Automate websites with an AI navigator
• Computer use preview — On macOS 15+, opt in to foreground desktop automation against the dev endpoint

macOS computer-use preview

This development-only preview is available only when the MCP server runs on macOS with YUTORI_ENV=dev. The runtime is Python-only and requires Python 3.10 or later. Setup installs CuaDriver, requests its macOS permissions, and prepares the optional native reasoning overlay:

uvx yutori-mcp computer-use setup
uvx yutori-mcp computer-use doctor
uvx yutori-mcp computer-use smoke

computer-use run executes one custom task from the terminal — the same run the MCP tool performs, with per-action progress printed as it happens:

uvx yutori-mcp computer-use run "In Calculator, compute 17 * 23 and report the result." --app Calculator

The run_computer_use_task tool controls the visible foreground desktop. Do not touch the Mac while it runs. Visible desktop content is sent to Yutori's dev model endpoint. Only one task can control a Mac at a time.

Workflow skills (for clients that support slash commands):

• /yutori-scout — Set up continuous web monitoring
• /yutori-research — Deep web research (async, 5–10 min)
• /yutori-browse — Browser automation
• /yutori-competitor-watch — Competitor monitoring template
• /yutori-api-monitor — API/changelog monitoring template

Installation

curl -LsSf https://astral.sh/uv/install.sh | sh

brew install uv

AI agent install (recommended)

Paste this into Claude Code, Codex, Cursor, Windsurf, or another coding agent:

Use https://yutori.com/api/llms.txt and set up Yutori for me.

Manual quick install

1. Run in terminal:

   uvx yutori-mcp login
   

   This will open Yutori Platform in your browser and save your API key locally.

   Or, manually add your API key

   Go to (https://platform.yutori.com) and add your key to the config file:

   mkdir -p ~/.yutori
   cat > ~/.yutori/config.json << 'EOF'
   {"api_key": "yt-your-api-key"}
   EOF
   

2. Install MCP using add-mcp (requires Node.js):

   npx add-mcp "uvx yutori-mcp"
   

   Pick the clients you want to configure.

3. Install workflow skills using skills.sh (requires Node.js):

   npx skills add yutori-ai/yutori-mcp -g
   

   Adds slash-command shortcuts like /yutori-scout, /yutori-research, and more.

   -g installs them at user scope. Omit -g if you want a project-local install instead.

   To list or remove skills later:

   npx skills ls -g
   npx skills remove -g yutori-login
   

4. Restart the tool you are using.

Manual per-client install

{
  "mcpServers": {
    "yutori": {
      "command": "uvx",
      "args": ["yutori-mcp"]
    }
  }
}

{
  "mcpServers": {
    "yutori": {
      "command": "uvx",
      "args": ["yutori-mcp"]
    }
  }
}

code --add-mcp '{"name":"yutori","command":"uvx","args":["yutori-mcp"]}'

{
  "mcpServers": {
    "yutori": {
      "command": "uvx",
      "args": ["yutori-mcp"]
    }
  }
}

{
  "mcp": {
    "allowed": ["yutori"]
  },
  "mcpServers": {
    "yutori": {
      "command": "uvx",
      "args": ["yutori-mcp"]
    }
  }
}

pip install yutori-mcp

Tools

See TOOLS.md for the full tool reference — Scout, Research, and Browsing tools with parameters, examples, and response formats.

Development

Setup

git clone https://github.com/yutori-ai/yutori-mcp
cd yutori-mcp
pip install -e ".[dev]"

Testing

pytest

Running locally

yutori-mcp login    # authenticate (one-time)
yutori-mcp          # run the server (or: python -m yutori_mcp.server)

Targeting the dev environment

The server hits the production API (https://api.yutori.com/v1) by default. For testing, point it at the dev stack (https://api.dev.yutori.com/v1) with the --env flag or the YUTORI_ENV environment variable (the flag wins if both are set):

yutori-mcp --env dev

Or in an MCP client config:

{
  "mcpServers": {
    "yutori-dev": {
      "command": "uvx",
      "args": ["yutori-mcp", "--env", "dev"]
    }
  }
}

Setting "env": {"YUTORI_ENV": "dev"} in the server config works too. An unknown environment name fails at startup rather than silently falling back to production. Plain login/logout/status commands manage production; pass --env dev to manage the separately stored dev credential described below. YUTORI_API_KEY overrides either stored credential when set.

Debugging with MCP Inspector

npx @modelcontextprotocol/inspector yutori-mcp

API Documentation

For full API documentation, visit docs.yutori.com.

License

Apache 2.0

Computer-use preview: authenticating against dev

login authenticates against production and saves a production key, which the dev stack rejects with a 401. Store a dev key separately:

uvx yutori-mcp --env dev login      # prompts for a key from platform.dev.yutori.com
uvx yutori-mcp --env dev status
uvx yutori-mcp --env dev logout

That writes an environments.dev entry alongside the existing top-level api_key, so one machine can hold both without either shadowing the other. YUTORI_API_KEY still takes precedence over both when set.

Computer-use preview: runtime dependency

The SDK-owned yutori.navigator.N2ComputerAgent and MacOSComputer provide the complete runtime. The MCP package pins SDK 0.9.1 and verifies the installed files against the immutable published wheel plus its packaged provenance during computer-use doctor. There is no TypeScript bundle, Node executable, or alternate harness. A normal source checkout needs no private dependency access:

uv sync --extra dev

SDK contributors may deliberately test an editable 0.9.1 checkout by setting YUTORI_MCP_ALLOW_EDITABLE_SDK=1; without that explicit override, doctor rejects editable or modified SDK installations.