USAspending MCP
A free, key-less MCP server for querying U.S. federal spending data from USAspending.gov, including contracts, grants, loans, recipients, agencies, and geographic breakdowns.
README
USAspending MCP
A free, key-less Model Context Protocol server for USAspending.gov — the U.S. government's official record of federal spending. Query contracts, grants, loans, recipients, agencies, and geographic breakdowns directly from the primary source. No API key, no account.
Built with FastMCP. Data covers FY2008-present.
Tools
| Tool | What it does |
|---|---|
usaspending_status |
Check API reachability and list supported award categories. |
list_agencies |
List/search federal toptier agencies; resolve an agency name to its code. |
search_awards |
Search awards (contracts/grants/loans/...) by recipient, agency, keyword, date. |
award_count |
Count matching awards broken down by type (contracts, grants, loans, idvs, ...). |
award_details |
Full detail for one award (amounts, period of performance, recipient, agencies). |
top_recipients |
Top recipients by dollars for an agency + award category over a period. |
spending_by_category |
Aggregate spending by recipient, agency, NAICS, PSC, CFDA, geography, etc. |
recipient_search |
Find recipients by name; returns UEI / DUNS and a recipient id. |
recipient_details |
Profile for one recipient: parent entity, business types, location, totals. |
spending_by_geography |
Spending by state / county / district, by recipient location or place of performance. |
Install
git clone https://github.com/jsconiers/usaspending-mcp.git
cd usaspending-mcp
python3 -m venv .venv
.venv/bin/pip install -r requirements.txt
.venv/bin/python test_usaspending.py # offline tests
Use with Claude Desktop
Add to claude_desktop_config.json (macOS:
~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"usaspending": {
"command": "/absolute/path/to/usaspending-mcp/.venv/bin/python",
"args": ["/absolute/path/to/usaspending-mcp/usaspending_mcp.py"]
}
}
}
Restart Claude Desktop, then ask things like "Who are the top DoD contractors this fiscal year?", "How much has Lockheed Martin been awarded?", or "Show federal contract spending by state."
Award categories
award_category accepts a friendly group that expands to USAspending award-type codes:
| Category | Codes |
|---|---|
contracts |
A, B, C, D |
idvs |
IDV_A ... IDV_E |
grants |
02, 03, 04, 05 |
direct_payments |
06, 10 |
loans |
07, 08 |
other |
09, 11, -1 |
all |
union of the above |
Pass award_type_codes explicitly to override.
Examples
list_agencies(search="defense") -> Department of Defense (code 097)
search_awards(recipient="Lockheed Martin") -> largest contracts + award_internal_id
award_count(recipient="Lockheed Martin") -> {contracts: 6501, grants: 21, idvs: 233, ...}
top_recipients(agency="Department of Defense") -> Lockheed, Electric Boat, Raytheon, ...
spending_by_category(category="naics", agency="Department of Defense")
spending_by_geography(geo_layer="state", agency="Department of Defense")
recipient_search(name="Lockheed Martin") -> UEI / DUNS + recipient_id
Typical flow: search_awards (or top_recipients) to find an award/recipient, then
award_details / recipient_details for the full record.
Data source
All data comes from the public USAspending.gov REST API (https://api.usaspending.gov/).
Please be considerate with request volume.
License
MIT — see LICENSE. Not affiliated with or endorsed by USAspending.gov or the U.S. government. For informational purposes only.
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。