Messari SDK TypeScript MCP Server

The MCP (Model Context Protocol) server enables AI assistants like Claude Code, Codex and Cursor to directly interact with Messari’s API services, allowing them to retrieve real-time crypto data and insights during conversations.

Key Benefits

AI Integration

Seamlessly connect AI assistants to Messari’s crypto data and services

Dynamic Tools

Discover and invoke API endpoints dynamically based on your needs

Flexible Configuration

Customize endpoint exposure and capabilities based on your use case

Comprehensive Coverage

Access the full range of Messari API services including Signal, Metrics, and AI

Quick Installation

You can run the MCP Server directly via npx:
export MESSARI_SDK_API_KEY="My API Key"
npx -y @messari/sdk-ts-mcp@latest

Using with MCP Clients

Many AI platforms, including Claude Code, Codex and Cursor, support the Model Context Protocol. You can configure your MCP client to use Messari’s MCP server:
{
  "mcpServers": {
    "messari_sdk_ts_api": {
      "command": "npx",
      "args": [
        "-y",
        "@messari/sdk-ts-mcp",
        "--client=claude",
        "--tools=dynamic"
      ],
      "env": {
        "MESSARI_SDK_API_KEY": "My API Key"
      }
    }
  }
}

Filtering for AI Endpoints Only

If you have a Pro or Lite tier API key and only want to access the AI completion endpoints, you can filter the available resources:
{
  "mcpServers": {
    "messari_sdk_ts_api": {
      "command": "npx",
      "args": [
        "-y",
        "@messari/sdk-ts-mcp@latest",
        "--client=cursor",
        "--resource",
        "'ai.*'"
      ],
      "env": {
        "MESSARI_SDK_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}
This configuration filters to only expose AI-related endpoints, which are available to Pro and Lite tier API keys at different rate limits. For more information on API tiers, permissions, and rate limits, see our API Permissions documentation.

Service Categories

The MCP server provides access to Messari’s comprehensive API services:
  • AI Toolkit - Chat completions, entity extraction, and market signals
  • Signal - Asset sentiment, mindshare, and trending topics
  • Metrics - Comprehensive crypto asset metrics and time series data
  • News - Crypto news feed and sources
  • Research - Analyst reports and insights
  • Token Unlocks - Vesting schedules and unlock events
  • Funding - Fundraising rounds, M&A, and investor data
  • User Management - Watchlists and API credits

Exposing Endpoints

The MCP server offers two approaches to expose endpoints:
  1. Individual Tools - Each API endpoint becomes a separate tool
  2. Dynamic Discovery - Three meta-tools that let the AI discover and invoke endpoints:
    • list_api_endpoints - Find available endpoints
    • get_api_endpoint_schema - Get detailed schema for an endpoint
    • invoke_api_endpoint - Execute any endpoint with parameters

Learn More

For comprehensive documentation, advanced configuration options, and all available tools, visit the MCP Server GitHub repository.