We've significantly expanded the capabilities of Messari's hosted MCP server. If you're using AI agents or tools that support the Model Context Protocol (MCP), you can now plug them into Messari's full intelligence layer, using your API key.

What's an MCP server?

The Model Context Protocol is a standard that lets AI assistants (Claude, Cursor, ChatGPT, and others) call external tools during a conversation. Messari's hosted MCP server gives these assistants direct access to our crypto data and research platform.

What's new

The MCP server originally shipped with a single tool, ask_messari, which lets agents ask natural language questions against Messari's data. We've now added five more tools that give agents structured, programmatic access:

Timeseries Data

• get_timeseries_catalog: Discover available datasets, metrics, and granularities from the Messari Timeseries API
• get_timeseries: Query timeseries data for crypto assets, protocols, and networks

Deep Research

• create_deep_research: Start an async deep research report on any crypto topic
• get_deep_research: Check status and retrieve a completed report
• list_deep_research: List all your deep research jobs

This means your agents can now do things like pull historical TVL data for a protocol, compare fee revenue across Layer-1 networks over time, or kick off the same deep research reports our analysts use, all without leaving your AI workflow.

Getting started

The hosted MCP server works with Claude (via Connectors), Claude Code, Cursor, ChatGPT, and any other MCP-compatible client. Setup takes about a minute.

👉 Read the docs 👉 Get your API key