mcp-server-omi: A OMI MCP server

Overview

A Model Context Protocol server for Omi interaction and automation. This server provides tools to read, search, and manipulate Memories and Conversations.

Tools

1. get_memories

   • Retrieve a list of user memories
   • Inputs:
     • limit (number, optional): Maximum number of memories to retrieve (default: 100)
     • categories (array of MemoryFilterOptions, optional): Categories of memories to retrieve (default: [])
   • Returns: JSON object containing list of memories

2. create_memory

   • Create a new memory
   • Inputs:
     • content (string): Content of the memory
     • category (MemoryFilterOptions): Category of the memory
   • Returns: Created memory object

3. delete_memory

   • Delete a memory by ID
   • Inputs:
     • memory_id (string): ID of the memory to delete
   • Returns: Status of the operation

4. edit_memory

   • Edit a memory's content
   • Inputs:
     • memory_id (string): ID of the memory to edit
     • content (string): New content for the memory
   • Returns: Status of the operation

5. get_conversations

   • Retrieve a list of user conversations
   • Inputs:
     • include_discarded (boolean, optional): Whether to include discarded conversations (default: false)
     • limit (number, optional): Maximum number of conversations to retrieve (default: 25)
   • Returns: List of conversation objects containing transcripts, timestamps, geolocation and structured summaries

Configuration

API Key

This local stdio server uses an MCP API key. In the current Omi macOS app, choose a key-based destination under Use omi memory anywhere and expand Manual installation to copy the generated key. In the cross-platform app, open Settings > Developer Settings, scroll to MCP Server, and create a key in the API Keys list. Copy the complete omi_mcp_... value when it is created.

The key can be provided with each tool call. If it is omitted, the server uses the OMI_API_KEY environment variable. The hosted Omi MCP endpoint also supports OAuth for registered cloud clients; this local stdio package uses the manual-key path.

Usage with Claude Desktop

Add this to your claude_desktop_config.json:

"mcpServers": {
  "omi": {
    "command": "uvx",
    "args": ["mcp-server-omi"],
    "env": {
      "OMI_API_KEY": "omi_mcp_YOUR_KEY_HERE"
    }
  }
}

"mcpServers": {
  "omi": {
    "command": "docker",
    "args": ["run", "--rm", "-i", "-e", "OMI_API_KEY=your_api_key_here", "omiai/mcp-server"]
  }
}

pip install mcp-server-omi

"mcpServers": {
  "omi": {
    "command": "mcp-server-omi",
    "env": {
      "OMI_API_KEY": "omi_mcp_YOUR_KEY_HERE"
    }
  }
}

Debugging

You can use the MCP inspector to debug the server. For uvx installations:

npx @modelcontextprotocol/inspector uvx mcp-server-omi

Or if you've installed the package in a specific directory or are developing on it:

cd path/to/servers/src/omi
npx @modelcontextprotocol/inspector uv run mcp-server-omi

To follow Claude Desktop logs while debugging:

# macOS
tail -n 20 -f ~/Library/Logs/Claude/mcp-server-omi.log

# Windows PowerShell
Get-Content "$env:APPDATA\Claude\logs\mcp-server-omi.log" -Tail 20 -Wait

Advanced

Custom Backend URL

If you are self-hosting the Omi backend, you can specify the API endpoint by setting the OMI_API_BASE_URL environment variable.

export OMI_API_BASE_URL="https://your-backend-url.com"

License

This MCP server is licensed under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more details, please see the LICENSE file in the project repository.