The Model Context Protocol (MCP) server lets AI assistants like Claude, Cursor, and other MCP clients call Vantafin data as tools - so your assistant can fetch quotes, fundamentals, filings, and more on its own, grounded in real data. This guide shows how to connect it. MCP is available on the Pro and Ultimate plans; you will need an API key.
What MCP gives your assistant
Instead of guessing, an MCP-connected assistant can retrieve current Vantafin data mid-conversation. That makes it ideal for:
- Research assistants that cite live filings and financials.
- Coding assistants (like Cursor) that pull data while you build.
- Any workflow where you want AI answers anchored to real market data.
Connect an MCP client
- Create an API key in Settings.
- In your MCP client, add Vantafin as an MCP server using the endpoint
https://mcp.vantafin.com/mcp. - Provide your API key as the credential (see your client's docs for where to put headers or tokens).
- Save, then confirm the Vantafin tools appear in the client.
The exact steps per client (Claude, Cursor, and others) are documented in the MCP server docs.
Example: using it in a conversation
Once connected, you can ask your assistant natural questions and it will call Vantafin tools to answer, for example:
- "Pull the latest quote and key fundamentals for a ticker."
- "Find recent SEC filings for a company and summarize the newest one."
- "Compare two companies on revenue growth."
MCP vs the in-app AI Chat
Vantafin also has a built-in AI Chat that is already grounded in Vantafin data, no setup required. Use MCP when you want that same data inside your own assistant or tools; use in-app AI Chat when you want the fastest path with zero configuration.
Troubleshooting
- Tools not appearing? Re-check the server URL and that your API key is valid and not revoked.
- Auth errors? Confirm the key is passed the way your client expects and that your plan includes MCP.
- Review your usage and limits in Settings.