home / mcp / usaspending mcp server
Provides MCP endpoints and tooling to query the USAspending.gov API for awards, agencies, and keyword searches.
Configuration
View docs{
"mcpServers": {
"flothjl-usaspending-mcp": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/flothjl/usaspending-mcp@main",
"usaspending-mcp"
]
}
}
}You can run the USAspending MCP server to enable AI agents to query and analyze USAspending.gov data through modular endpoints. It provides focused tools and a lightweight runtime that makes it easy to fetch agency awards, specific award details, agency lists, and keyword-based searches for a given year.
Start by launching the MCP server using its runtime configuration. You will run a local MCP client or orchestration layer that speaks the MCP protocol and connects to this server to perform the available actions. Once running, you can perform actions such as getting awards by agency and year, retrieving information about a particular award by its generated ID, performing keyword searches for a year, and pulling a list of agencies necessary for other calls.
Prerequisites: you need a runtime capable of MCP protocol interactions. In this setup, you will run a local stdio MCP server using the provided command.
extensions:
usaspending-mcp:
args:
- --from
- git+https://github.com/flothjl/usaspending-mcp@main
- usaspending-mcp
cmd: uvx
enabled: true
envs: {}
name: usaspending
type: stdio
GOOSE_MODEL: gpt-4o-mini
GOOSE_PROVIDER: openaiConfiguration, startup flow, and environment details help you tailor how the server runs in your environment.
Security and usage notes: keep any API-related endpoints secured, and follow best practices for handling API keys and access controls when integrating with an MCP client.
Search for awards for a given agency id for a given year.
Get information about a given award. Requires the generated unique award ID.
Broad search for given keywords for a given year.
Get a list of all agencies and associated data needed for other calls (i.e. agency id).