All FDIC BankFind endpoints are implemented except `/financials`, whose schema is large enough to break code generation — that is stated up front rather than discovered when a call fails. The shared parameter set is what makes the tools usable in practice: the same filter, field-selection and paging arguments work everywhere, so one query pattern carries across institutions, locations and failures.
A Rust MCP server over the FDIC BankFind APIs, shipped as a public container image. It brings structured US banking data — every insured institution, its branches, its history, and the failure record — into an assistant without an API integration project.
- Institution demographics with a free-text search parameter, and branch locations — `get_institutions`, `get_locations`
- The historical bank failure record, with aggregation parameters for grouping and summing — `get_failures`
- Structure change events over time, searchable and aggregatable — `get_history`
- Summary of Deposits data and historical aggregates by year — `get_sod`, `get_summary`
- Demographic summaries — `get_demographics`
- Shared parameters on every call: `filters` in FDIC BankFind syntax, `fields` to choose columns, `limit` and `offset` for paging, `sort_by` and `sort_order`, and `file_format` for json, csv or xml
Docker, running. The FDIC API key is optional — pass it as `api_key` if you have one. Run `docker run -i --rm ghcr.io/clafollett/fdic-bank-find-mcp-server:main`; the `-i` flag is required for stdio. Building from source instead needs the Rust toolchain and `cargo build --release`.
One command — docker run -i --rm ghcr.io/clafollett/fdic-bank-find-mcp-server:main