FDIC BankFind MCP Server
Provides programmatic access to the FDIC BankFind Suite API, enabling users to query public data on FDIC-insured financial institutions, bank failures, and branch locations. It supports advanced filtering for financial reports, demographics, and institutional history without requiring an API key.
README
FDIC BankFind MCP Server
fdic-mcp-server is an MCP (Model Context Protocol) server for the FDIC BankFind Suite API. It gives LLM hosts a clean way to search FDIC-insured institutions, retrieve public banking records, and run common multi-bank analysis workflows without custom FDIC API plumbing.
It is useful when you want an MCP-compatible client to answer questions about banks, failures, branches, quarterly financials, deposit data, or peer performance using a stable tool surface and machine-readable responses.
Table Of Contents
- Project Status
- Why This Project Exists
- Documentation
- Installation
- Usage
- Available Tools
- Data Notes
- Support
- Contributing
- License
Project Status
Active development. The server is usable today and the tool surface is covered by tests, but the project is still evolving as client support and analysis workflows improve.
Why This Project Exists
The FDIC BankFind Suite API is public and useful, but it is not packaged for MCP clients out of the box. This project solves that by:
- exposing BankFind datasets as MCP tools
- preserving both human-readable and machine-readable responses
- adding server-side analysis helpers for multi-bank comparison workflows
- supporting both local stdio hosts and remote HTTP hosts
Documentation
- GitHub Pages docs entry point: https://jflamb.github.io/fdic-mcp-server/
- Hosted MCP endpoint: https://bankfind.jflamb.com/mcp
- Local docs home: docs/index.md
- Getting started: docs/getting-started.md
- Prompting guide: docs/prompting-guide.md
- Usage examples: docs/usage-examples.md
- Tool reference: docs/tool-reference.md
- Client setup: docs/clients.md
- Troubleshooting and FAQ: docs/troubleshooting.md
- Compatibility matrix: docs/compatibility-matrix.md
- Technical specification: docs/technical/specification.md
- Architecture: docs/technical/architecture.md
- Key decisions: docs/technical/decisions.md
- Release notes: docs/release-notes/index.md
- Security policy: SECURITY.md
Installation
Prerequisites:
- Node.js 18 or later
- npm
Run directly without a global install:
npx fdic-mcp-server
Install globally:
npm install -g fdic-mcp-server
fdic-mcp-server
Install from GitHub Packages:
npm install -g @jflamb/fdic-mcp-server --registry=https://npm.pkg.github.com
fdic-mcp-server
Install from source:
git clone https://github.com/jflamb/fdic-mcp-server.git
cd fdic-mcp-server
npm install
npm run build
Usage
Hosted Endpoint
If your MCP host supports remote MCP URLs, use:
https://bankfind.jflamb.com/mcp
Run Locally
Stdio transport:
node dist/index.js
HTTP transport:
TRANSPORT=http PORT=3000 node dist/index.js
The HTTP MCP endpoint is http://localhost:3000/mcp.
Minimal MCP Configuration
{
"mcpServers": {
"fdic": {
"command": "npx",
"args": ["-y", "fdic-mcp-server"]
}
}
}
If you are running from a local clone instead of the published package:
{
"mcpServers": {
"fdic": {
"command": "node",
"args": ["/path/to/fdic-mcp-server/dist/index.js"]
}
}
}
Client-specific setup details are in docs/clients.md.
Usage Examples
Find active banks in North Carolina with assets over $1 billion:
filters: STNAME:"North Carolina" AND ACTIVE:1 AND ASSET:[1000000 TO *]
Get the 10 costliest bank failures since January 1, 2000:
filters: FAILDATE:[2000-01-01 TO *]
sort_by: COST
sort_order: DESC
limit: 10
Compare North Carolina banks between two quarterly report dates:
state: North Carolina
start_repdte: 20211231
end_repdte: 20250630
sort_by: asset_growth
sort_order: DESC
(fdic_compare_bank_snapshots)
Build a peer group for a specific bank:
cert: 29846
repdte: 20241231
(fdic_peer_group_analysis)
More examples are in docs/usage-examples.md.
Available Tools
| Tool | Description |
|---|---|
fdic_search_institutions |
Search FDIC-insured banks and savings institutions |
fdic_get_institution |
Get details for a specific institution by CERT number |
fdic_search_failures |
Search failed bank records |
fdic_get_institution_failure |
Get failure details for a specific institution |
fdic_search_locations |
Search branch and office locations |
fdic_search_history |
Search structural change events such as mergers and name changes |
fdic_search_financials |
Search quarterly Call Report financial data |
fdic_search_summary |
Search annual financial summary data |
fdic_search_sod |
Search Summary of Deposits branch-level deposit data |
fdic_search_demographics |
Search quarterly demographics and market-structure data |
fdic_compare_bank_snapshots |
Compare two reporting snapshots across banks and rank growth and profitability changes |
fdic_peer_group_analysis |
Build a peer group and rank an institution against peers on financial metrics |
Two tools are server-side analysis helpers:
fdic_compare_bank_snapshotsbatches roster lookup, financial snapshots, and optional demographics snapshots inside the MCP serverfdic_peer_group_analysisbuilds a peer group from asset size, charter class, and geography criteria and then ranks an institution against peers
Data Notes
- Monetary values are generally reported in thousands of dollars.
CERTis the stable FDIC institution identifier.- Financial and demographics datasets are quarterly and use
REPDTEinYYYYMMDD. - Summary data is annual and uses
YEAR. - SOD data is annual branch-level data as of June 30.
- Do not mix quarterly financial data and annual branch data without stating the date basis.
Support
Use the GitHub issue tracker for bugs, documentation problems, and feature requests: https://github.com/jflamb/fdic-mcp-server/issues
The main support docs are:
Contributing
Contributor guidance lives in CONTRIBUTING.md.
For local validation, run:
npm run typecheck
npm test
npm run build
License
This project is licensed under the MIT License.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。