USAspending (Federal Spending) MCP for AI. Map money flow across agencies, states, and years.

USAspending MCP Server - Federal Spending Analysis

Listen up. This server connects you directly to the official USAspending database. Your agent uses it to pull real-time financial data showing exactly how the U.S. government spends its cash. You're not just looking at totals; you're tracking every single dollar, from major department budgets down to a specific transaction record.

It gives you deep visibility into federal money flows.

Agency Deep Dives and Budget Status
You wanna know what an entire department's up to? You pull a general summary page for any major agency using get_agency_overview. For the full picture, your agent can retrieve detailed budgetary resources and current obligation amounts across multiple fiscal years via get_agency_budgetary_resources, or it can summarize all transactions tied to one specific federal group with get_agency_awards.

If you're looking at a 'top tier' department, get_toptier_agencies gives you the key metrics for all designated government giants. To drill down further, you can use get_agency_sub_agencies to list every internal office or sub-agency under that big department and see exactly how much money they've obligated. You also get a count of different award types an agency managed in a given year using get_agency_awards_count.

For agencies handling emergency situations, you can run dedicated checks: get_disaster_overview gives broad insights into general disaster funding, and get_disaster_agency_spending focuses on specific spending related to those emergencies. You'll also get the aggregated obligation and outlay amounts for awards issued after a crisis using get_disaster_award_amount.

Tracking Specific Awards and Money Flows
You can track everything that happens with any single contract or grant. Use get_award to pull all the detailed info on one federal award record, including its original funding source via get_award_funding. You wanna see what kind of money it is? get_award_types maps out different categories of awards for you.

If an award has sub-contracts, you use get_subawards to list those details that fall under the main parent agreement. To get the granular breakdown, your agent lists specific financial transactions against a parent award using get_transactions. You can also check how many new awards came in over time by running search_new_awards_over_time, or you can find out which agency awarded an amount of money at a glance with get_agency_awards.

Filtering Data Across Scope, Location, and Time
The data is massive, so you gotta filter it down. You don't wanna look at everything; you want to know where the cash went. Use search_spending_by_geography to find and filter spending amounts specifically by state, county, or congressional district—it pinpoints exactly where money landed.

For a historical view, run search_spending_over_time to aggregate dollar amounts across different years. You can also search for federal awards using multiple criteria with search_spending_by_award, or group all spending data by category, agency type, or recipient type for visualization purposes via search_spending_by_category. When you need to narrow down the recipients, your agent uses autocomplete_recipient and then pulls basic details about them with get_recipient, or gets deep info on their corporate structure using get_recipient_children.

Data Utility and Downloads
If you're working with millions of records, you gotta get the data out. Your agent can generate a single ZIP file download covering awards, subawards, and transactions for any search query via download_search. If you just need one type of record, you can make a big package containing all selected award data—including subawards—and zip it up in CSV format using bulk_download_awards.

You can also check on those large downloads anytime with get_download_status. Need to autocomplete something? Use autocomplete_awarding_agency for agency names, or autocomplete_location when searching by geography. For definitions, the server gives you get_glossary and autocomplete_glossary to make sure you're using the right terminology; it also lists all available glossary terms via get_data_dictionary.

Finally, if you just wanna see who's in the system, list_recipients returns the full list of every entity recorded as a recipient.