Agents & MCP

Connect AI assistants to your Lighthouse portfolio data using the Model Context Protocol

The Lighthouse MCP server is currently in beta. The available tools and their responses may change as we iterate on this feature. We welcome your feedback.

What is the MCP server?

The Model Context Protocol (MCP) is an open standard that allows AI assistants to connect to external data sources and tools. Lighthouse exposes an MCP server that gives your AI assistant read-only access to your portfolio data.

This means you can ask your AI assistant natural language questions about your portfolios, positions, performance, strategies, and more — and it will fetch live data from Lighthouse to answer them.

What can your AI assistant do?

Once connected, your AI assistant can use the following tools:

Workspaces & Portfolios

• List all your workspaces
• List and inspect portfolios across workspaces
• Check portfolio sync status

Accounts

• List all accounts (wallets, exchanges, manual) in a portfolio
• Get details for a specific account

Snapshots & Positions

• List snapshots (point-in-time portfolio states)
• Get the latest snapshot with full position details
• Get a flattened view of all positions with network, platform, symbol, amount, and USD value
• View snapshot history over time

Performance & Prices

• Get PnL charts, value charts, asset breakdown, and top gainers/losers
• Get current prices for your top portfolio assets

Strategies

• List all trading strategies with PnL and deployed amounts
• Get detailed strategy data including logs and history

Configure your AI client

The Lighthouse MCP server is available at:

https://lighthouse.one/mcp

It uses OAuth 2.0 for authentication. Your AI client will open a browser window for you to log in on first use — no API keys or tokens to manage manually. The MCP server respects your Lighthouse account permissions — you can only access the workspaces and portfolios you have access to in Lighthouse.

Claude Code

Run the following command in your terminal:

claude mcp add lighthouse --transport http https://lighthouse.one/mcp

The next time Claude Code calls a Lighthouse tool, it will open your browser to complete the OAuth login.

For more details, see the Claude Code MCP documentation.

Claude Desktop

1. Open Connectors settings

   Open Claude Desktop and navigate to Customize > Connectors.

2. Add a custom connector

   Click the + button, then select Add custom connector.

3. Configure the connector

   Enter a name (e.g. Lighthouse) and the MCP server URL:

   https://lighthouse.one/mcp
   
4. Connect

   Click Add, then click Connect on the newly added connector to complete the OAuth login with your Lighthouse account.

5. Set permissions

   For a smoother experience, set the connector’s permissions to Always Allow. This is safe to do as all Lighthouse MCP tools are read-only.

For more details, see the Claude custom connectors documentation.

Claude Web

1. Open Connectors settings

   Navigate to claude.ai/customize/connectors.

2. Add a custom connector

   Click the + button, then select Add custom connector.

3. Configure the connector

   Enter a name (e.g. Lighthouse) and the MCP server URL:

   https://lighthouse.one/mcp
   
4. Connect

   Click Add, then click Connect on the newly added connector to complete the OAuth login with your Lighthouse account.

5. Set permissions

   For a smoother experience, set the connector’s permissions to Always Allow. This is safe to do as all Lighthouse MCP tools are read-only.

For more details, see the Claude custom connectors documentation.

OpenCode

Add the following to your opencode.json configuration file:

{
  "mcp": {
    "lighthouse": {
      "type": "remote",
      "url": "https://lighthouse.one/mcp"
    }
  }
}

For more details, see the OpenCode MCP documentation.

Cursor

Add a .cursor/mcp.json file to your project root (or configure it globally):

{
  "mcpServers": {
    "lighthouse": {
      "url": "https://lighthouse.one/mcp"
    }
  }
}

For more details, see the Cursor MCP documentation.

Example prompts

Once connected, try asking your AI assistant:

• “Show me my portfolio performance for the last 30 days”
• “What are my largest positions by USD value?”
• “List all my strategies and their PnL”
• “Which assets gained the most this week?”
• “Give me a breakdown of my portfolio across chains”