The three tools around queryVCData are what make it work: schema, per-table detail and worked examples mean you can ask about recent Series A fintech deals without handing the model a data dictionary first. An active Fundable subscription is what makes it return anything.
A remote, hosted MCP server over Fundable's proprietary venture-capital dataset, covering funding rounds, investors, valuations, founders and firms. There is nothing to install or run locally.
- A read-only query run against the dataset, returning the rows themselves (queryVCData)
- The dataset's schema, business rules and query guidelines, so a model writes queries that fit it (getDatasetContext)
- The table list, and per table the column names, types and constraints (listDatasetTables, getTableDetails)
- Worked example queries by category, covering funding, investors, people, industries and geography (getQueryExamples)
A Fundable account with an active subscription, signed in over OAuth. Point a client at https://fundable-mcp.jacob-57a.workers.dev/mcp over Streamable HTTP; older clients bridge to it with mcp-remote.
One command — npx mcp-remote https://fundable-mcp.jacob-57a.workers.dev/mcp
