FDIC BankFind MCP Server: U.S. Banking API
Bring FDIC BankFind data to your AI tools with this MCP server, delivering structured U.S. banking APIs for workflows and integrations.
npx -y @clafollett/fdic-bank-find-mcp-serverOverview
This MCP (Model Context Protocol) server exposes FDIC BankFind information as a structured, machine-readable API that can be embedded inside AI tools, chatbots, and automation pipelines. It translates FDIC BankFind data into a predictable set of endpoints and tool descriptors compatible with the MCP pattern so language models and other agents can request precise bank details, perform searches, and retrieve metadata without scraping HTML pages.
By serving data in a consistent JSON schema and optionally caching FDIC responses, the server reduces latency and simplifies integration of U.S. banking information into workflows such as customer support assistants, compliance checks, document enrichment, and financial data pipelines.
Features
- MCP-compatible endpoints and tool metadata for easy AI integration
- Search banks by name, routing number, FDIC certificate, city/state, and ZIP
- Retrieve detailed bank profiles (address, status, certificate, files, and more)
- Optional caching to reduce FDIC API calls and improve throughput
- Lightweight Node.js server with Docker support
- CORS enabled for browser-based tooling and web clients
- Simple OpenAPI or JSON schema export for tool discovery (when enabled)
Installation / Configuration
Prerequisites: Node.js (16+), npm or yarn. Optionally Docker.
Clone and install:
Environment configuration (example .env):
# .env
PORT=3000
FDIC_API_KEY=your_fdic_api_key_or_empty_if_unneeded
FDIC_BASE_URL=https://banks.data.fdic.gov/api
CACHE_TTL_SECONDS=3600
LOG_LEVEL=info
Start locally:
# or for development with auto-reload
Run with Docker:
# build image
# run container
If the FDIC API requires no key in your environment, leave FDIC_API_KEY blank and the server will default to the public FDIC endpoint or to a local cache, depending on configuration.
Available Resources
- GitHub repository: https://github.com/clafollett/fdic-bank-find-mcp-server
- MCP endpoints (examples described below)
- OpenAPI/JSON Schema export (if enabled) for client generation and validation
Example primary endpoints (default server prefix: /mcp):
| Endpoint | Method | Purpose |
|---|---|---|
| /mcp/describe | GET | Returns MCP tool descriptor(s) for model/tool discovery |
| /mcp/search | GET | Search banks by name, city, state, ZIP, routing, certificate |
| /mcp/institution/{cert} | GET | Full profile for an institution by FDIC certificate number |
| /mcp/routing/{routingNumber} | GET | Lookup bank by routing number |
| /openapi.json | GET | Optional OpenAPI spec for the server |
Example: query the MCP search
Sample response (trimmed):
Use Cases
- AI-powered customer support: Provide a virtual agent with authoritative bank information to answer customer questions about branch locations, insurance status, or routing numbers.
- Compliance and onboarding automation: Validate a bank’s FDIC certificate or routing number during account opening flows and enrich customer records with normalized bank metadata.
- Document enrichment: Extract an institution name from a document and use the MCP endpoint to append structured bank details (address, status, identifiers) for downstream processing.
- Monitoring and alerts: Periodically query the MCP server for status changes (e.g., mergers, closures) and trigger notifications or workflow actions when a monitored institution changes state.
- Prompt tooling and chaining with LLMs: Provide the MCP tool descriptor to an LLM so it can call the precise endpoint rather than hallucinating bank facts.
Integration Tips
- Use the /mcp/describe endpoint to programmatically retrieve tool metadata for runtime registration with an agent framework or LLM orchestration layer.
- Cache results when possible: routing numbers and certificate metadata rarely change and are appropriate for longer TTLs to reduce external API usage.
- Sanitize and validate inputs: although the server normalizes queries, validate user-provided routing numbers and certificate numbers to avoid unnecessary calls.
- Rate limiting: if you deploy at scale, add an API gateway or cloud rate limiting layer to avoid exceeding FDIC API quotas.
For full implementation details, API schema, and example client code, see the project repository: https://github.com/clafollett/fdic-bank-find-mcp-server.