fec-mcp
An MCP server that searches official FEC PDF rulebooks for compliance and contribution limits, and provides real-time lookups against the OpenFEC API for candidates, committees, filings, and financial data.
README
fec-mcp
An MCP (Model Context Protocol) server for Federal Election Commission rules, regulations, and data covering PACs, party committees, and candidates.
It provides two complementary tool families:
- Rulebook search -- full-text search over the FEC's own official PDF guides (campaign guides for candidates, party committees, and PACs, plus the contribution-limits chart). This is the authoritative source for compliance questions and dollar limits: results are quoted directly from the PDFs with filename + page citations, not hardcoded or model-recalled figures.
- Live OpenFEC data -- real-time lookups against the public OpenFEC API for candidates, committees, filings, financial totals, elections, and the reporting calendar.
Setup
1. Install
python3 -m venv .venv
source .venv/bin/activate
pip install -e .
2. Add the FEC rulebook PDFs
Place the official FEC PDF guides in data/rulebooks/, for example:
- The FEC's Campaign Guide for Congressional Candidates and Committees
- The FEC's Campaign Guide for Political Party Committees
- The FEC's Campaign Guide for Nonconnected Committees (PACs)
- The current Contribution Limits chart
These are published on fec.gov under its Campaign Guides / Contribution Limits resources; grab the current versions from there (this repo does not bundle download links since they change over time -- verify you have the current cycle's documents).
Any PDF you drop in data/rulebooks/ is picked up automatically -- the
server builds a search index (cached in data/rulebooks/.index/) the first
time it starts and rebuilds it whenever files are added, removed, or
changed.
3. Get an OpenFEC API key (optional but recommended)
Live OpenFEC lookups work out of the box with the shared DEMO_KEY, which
is heavily rate-limited. For real use, get a free key at
https://api.data.gov/signup/ and set it:
export FEC_API_KEY=your_key_here
4. Configure your MCP client
For Claude Code / Claude Desktop, add to your MCP config:
{
"mcpServers": {
"fec": {
"command": "/absolute/path/to/FEC_MCP/.venv/bin/fec-mcp",
"env": {
"FEC_API_KEY": "your_key_here"
}
}
}
}
Or run directly for local testing:
fec-mcp
Tools
Rulebook search (grounded in the PDFs you add):
| Tool | Purpose |
|---|---|
list_rulebook_sources |
List loaded PDFs and page counts |
search_rulebooks |
Full-text search across all loaded PDFs |
get_rulebook_page |
Read the full text of one page |
Live OpenFEC data:
| Tool | Purpose |
|---|---|
search_candidates / get_candidate / get_candidate_totals |
Candidate lookup and financials |
search_committees / get_committee / get_committee_filings / get_committee_totals |
PAC/party/campaign committee lookup, filings, financials |
search_filings |
Cross-committee filing search |
search_elections |
Election search by state/office/cycle |
get_reporting_calendar |
FEC reporting and election deadline dates |
Notes on scope
- No hardcoded contribution limits or rule text. OpenFEC's live API has no endpoint for contribution limits or regulation text, and those figures change (limits are inflation-adjusted every two-year cycle) -- baking them into code risks going stale silently. Instead, this server searches the FEC's own PDFs you provide, so answers are always sourced to a specific document and page.
- Not legal advice. Tool output should be treated as a research aid and cited back to its source (PDF page or OpenFEC record), not relied on as a substitute for consulting the FEC or legal counsel.
Development
pip install -e ".[dev]"
pytest
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。