API Key: Go to your workspace → Settings → Agent Keys → Create new key
Workspace ID: Copy from the URL when viewing your workspace (/w/<workspace-id>)

2. Add to your MCP config

Claude Desktop (claude_desktop_config.json):

{
  "mcpServers": {
    "teammem": {
      "command": "npx",
      "args": ["-y", "teammem-mcp"],
      "env": {
        "TEAMMEM_API_KEY": "tm_sk_your_key_here",
        "TEAMMEM_WORKSPACE": "your-workspace-id",
        "TEAMMEM_API_URL": "https://your-api.vercel.app"
      }
    }
  }
}

Cursor (.cursor/mcp.json):

{
  "mcpServers": {
    "teammem": {
      "command": "npx",
      "args": ["-y", "teammem-mcp"],
      "env": {
        "TEAMMEM_API_KEY": "tm_sk_your_key_here",
        "TEAMMEM_WORKSPACE": "your-workspace-id",
        "TEAMMEM_API_URL": "https://your-api.vercel.app"
      }
    }
  }
}

3. Start using it

Your AI tool now has access to these tools:

Tool	What it does
`search`	Semantic + keyword search across all entries
`read_entry`	Read full content of a specific entry
`write_entry`	Create new entries (meeting notes, decisions, contacts)
`update_entry`	Update existing entries with conflict detection
`delete_entry`	Delete an entry
`list_collections`	See all collections in the workspace
`query_structured`	Filter/sort structured data (like SQL for your knowledge base)
`workspace_info`	Get workspace overview and collection summary

Environment Variables

Variable	Required	Description
`TEAMMEM_API_KEY`	Yes	Agent API key (starts with `tm_sk_`)
`TEAMMEM_WORKSPACE`	Yes	Workspace UUID
`TEAMMEM_API_URL`	No	API base URL (default: `http://localhost:3001`)

SSE Transport (Remote)

For remote connections without running a local process:

teammem-mcp-sse

Or configure your MCP client to connect via SSE:

{
  "mcpServers": {
    "teammem": {
      "transport": "sse",
      "url": "https://your-mcp-server.com/sse",
      "headers": {
        "Authorization": "Bearer tm_sk_your_key_here"
      }
    }
  }
}