Yutu MCP Server & CLI for YouTube Automation

Overview

Yutu MCP Server & CLI is a developer-focused toolkit for automating YouTube workflows. It exposes a Model Context Protocol (MCP) compatible server plus a command-line interface that together let you schedule uploads, manage channel assets, and surface analytics in programmatic flows. The MCP server acts as a bridge between automated agents (like LLMs) and YouTube operations, while the CLI provides a direct developer experience for scripting and interactive use.

This combination is useful when you want repeatable, auditable YouTube automation: scheduled publishing, consistent metadata updates, bulk operations across channels, and automated analytics collection for dashboards or downstream ML pipelines. The design separates tool execution (MCP) from orchestration (agents, cron jobs, CI), so you can integrate Yutu into a variety of systems and workflows.

Features

• MCP-compatible server that exposes YouTube operations as callable tools
• CLI for single-command operations and scripting
• OAuth and API-key integration with YouTube Data API
• Scheduling primitives for timed uploads and publish windows
• Tools for uploads, metadata updates, playlist management, and comments moderation
• Analytics retrieval endpoints for views, watch time, and engagement metrics
• Webhook and task-execution hooks for integration with CI/CD, cron, or agent frameworks
• Local development mode and Docker support for deployment

Installation / Configuration

Prerequisites:

• Node.js 18+ (or compatible runtime)
• YouTube Data API credentials (OAuth client ID/secret or API key)
• (Optional) Docker

Clone and install:

git clone https://github.com/eat-pray-ai/yutu.git
cd yutu
npm install

Create a .env file (example):

YOUTUBE_CLIENT_ID=your-client-id.apps.googleusercontent.com
YOUTUBE_CLIENT_SECRET=your-client-secret
YOUTUBE_API_KEY=your-api-key
OAUTH_REDIRECT_URI=http://localhost:3000/oauth2callback
MCP_PORT=8080
CLI_DEFAULT_SERVER=http://localhost:8080

Run in development:

npm run dev
# or build + run
npm run build
npm start

Docker (example):

docker build -t yutu-mcp .
docker run -p 8080:8080 --env-file .env yutu-mcp

CLI installation (local/project):

# if included as an npm package in the repo
npm link
# or run via npx / node directly
npx yutu-cli --help

Available Tools / Resources

The MCP server exposes a set of tools (examples below). These become callable actions for agents or the CLI.

Resources:

• Source and issues: https://github.com/eat-pray-ai/yutu
• YouTube Data API docs: https://developers.google.com/youtube/v3
• MCP spec (if applicable): consult project docs for protocol details

Use Cases

1. Schedule a weekly episode series

   • Use the CLI to upload a video and schedule it:

     yutu upload --file ./episode1.mp4 --title "Weekly Show #1" --schedule "2026-05-01T12:00:00Z"
     

   • Or have an LLM agent call the MCP upload_video and schedule_publish tools to automate recurring episodes.

2. Bulk-update metadata across videos

   • Query videos via get_analytics or list endpoint, then call update_metadata in a loop to add consistent tags or improve descriptions before a marketing campaign.

3. Analytics-driven content iteration

   • Periodically run get_analytics to extract watch time and retention metrics, store results in your analytics DB, and trigger notifications when performance drops below thresholds.

4. Automated moderation and compliance

   • Use moderate_comments to flag or remove comments that match policy rules; wire a webhook to notify moderators when escalations occur.

5. CI/CD content pipeline

   • Integrate the MCP server with CI pipelines to auto-publish builds of screencasts, release notes, or changelog videos after a successful release job.

Example CLI Commands

# Upload and schedule
yutu upload --file ./video.mp4 --title "New Video" --description "Auto-upload" --tags "ai,automation" --schedule "2026-05-01T12:00:00Z"

# List channels
yutu channels list

# Fetch analytics for a video
yutu analytics get --video-id abc123 --range "30d"

# Update metadata
yutu metadata update --video-id abc123 --title "Updated Title" --description "Updated desc"

Next steps

• Configure OAuth credentials and perform the initial OAuth flow to authorize the service account or user.
• Review and extend tools to fit your operational policies (rate limits, retry behavior, audit logs).
• Combine with agent frameworks (LLMs, schedulers) to build higher-level automation workflows.