MCP Setup

Connect BeatOps to your AI agent

What is MCP?

MCP (Model Context Protocol) lets AI tools like Claude and Cursor interact with other software. When you connect BeatOps via MCP, your AI assistant can list your beats, analyze audio, generate assets, and upload to platforms - all through conversation.

Instead of clicking through menus, just tell your AI: "Upload my latest beat to YouTube and SoundCloud."

Prerequisites

  • BeatOps desktop app installed and running
  • An MCP-compatible AI tool (Claude Desktop, Claude Code, Cursor, etc.)

The CLI (beatops) and MCP server (beatops-mcp) are included with the desktop app - no extra installation needed.

Share this page with your AI agent

The sections below are written for developers and AI agents. If you are setting up BeatOps for an AI tool, share this page URL with your agent - it can follow these instructions directly.

Configuration

Claude Desktop

Add the following to your claude_desktop_config.json:

{
  "mcpServers": {
    "beatops": {
      "command": "beatops-mcp"
    }
  }
}

Claude Code

Add the following to your project's .mcp.json:

{
  "mcpServers": {
    "beatops": {
      "command": "beatops-mcp",
      "type": "stdio"
    }
  }
}

Other MCP Clients

The MCP server runs via stdio transport. Point your MCP client to the beatops-mcp command.

Available Tools

Safe (no confirmation needed)

These tools run without confirmation:

  • list_beats - List all beats in the library
  • get_beat(beat_name) - Get detailed info for a specific beat
  • get_platform_status - Check YouTube and SoundCloud connection status
  • get_readiness(beat_name) - Check if a beat is ready to upload
  • render_template(beat_name, platform) - Preview rendered metadata for a platform
  • analyze_beat(beat_name) - Analyze audio to detect BPM, key, genre, and mood
  • generate_images(beat_name) - Generate thumbnails and cover art
  • generate_video(beat_name) - Generate the upload video
  • generate_tags(beat_name) - Generate SEO tags

Confirm (requires approval)

These tools ask for your approval before executing:

  • upload_beat(beat_name, platform) - Upload to YouTube or SoundCloud

Permission model: Most tools run without confirmation. Only uploads to YouTube and SoundCloud require your approval. Dangerous operations (deleting beats, modifying credentials) are blocked entirely.

Example Conversation

Here's what it looks like when your AI agent handles a full upload:

Claude + BeatOps MCP
Y
You
Upload my latest beat to YouTube and SoundCloud
C
Claude
list_beats
analyze_beat("Midnight")
generate_images("Midnight")
generate_video("Midnight")
generate_tags("Midnight")
Found and prepared Midnight (96 BPM, G minor, Soulful Boom Bap). Thumbnails, video, and tags are ready. Want me to upload?
Y
You
Yes, go ahead
C
Claude
upload_beat("Midnight", "youtube")
upload_beat("Midnight", "soundcloud")
Done! Your beat is live on both platforms.
youtube.com/watch?v=abc123 soundcloud.com/producer/midnight

Tool Reference

The individual tools your AI agent uses in these workflows:

Upload a beat

  1. get_beat("Midnight") - Get beat details
  2. get_readiness("Midnight") - Verify the beat is ready
  3. upload_beat("Midnight", "youtube") - Upload to YouTube

Full pipeline

  1. list_beats - Find your beat
  2. analyze_beat("Midnight") - Detect BPM, key, genre, mood
  3. generate_images("Midnight") - Create thumbnails
  4. generate_video("Midnight") - Create upload video
  5. generate_tags("Midnight") - Generate SEO tags
  6. upload_beat("Midnight", "youtube") - Upload

Check status

  1. get_platform_status - See which platforms are connected
  2. list_beats - See all beats in your library
← Back to Help Center