AwardCast MCP¶
Remote MCP server (Streamable HTTP, no auth) serving public US federal contracting data to AI agents. Eight tools cover open solicitations from SAM.gov, buying profiles for federal agencies, award history for contractors, contracts heading to recompete with the incumbent, historical closing-price ranges by agency and NAICS code, and a forecast accuracy scorecard. Data comes from USASpending/FPDS and SAM.gov.
Server type: Remote (Streamable HTTP)
Auth: None (public data)
Endpoint: https://awardcast.ai/mcp
Tools: 8 (search, fetch, solicitations, buying profiles, award history, recompete radar, price ranges, forecast scorecard)
Pricing: Free
Category: Government Contracting Data
Built by: AwardCast (awardcast.ai), repo ChosingDept/awardcast-mcp
Why This Matters for Operators¶
Federal contracting is a data war. Incumbents defend renewals, challengers hunt recompetes, and every player needs to know what agencies actually pay for a given NAICS code. That research normally means sitting inside SAM.gov and USASpending with their clunky interfaces. AwardCast compresses it into eight tools an agent can call directly.
The search and fetch tools follow the deep-research contract: results carry citable URLs, so a bid team gets an evidence trail, not a summary. The recompete radar is the standout: contracts approaching recompete are flagged with the incumbent named, which converts "we should look at gov work" into a concrete target list.
Tools & Capabilities¶
| Capability | What it does |
|---|---|
search / fetch |
Deep-research style queries with citable source URLs on every result |
| Open solicitations | Active SAM.gov opportunities, filterable for pipeline building |
| Agency buying profiles | How a given federal agency buys: volumes, NAICS mix, preferred vehicles |
| Award history | Historical awards by contractor, for incumbent analysis and past-performance research |
| Recompete radar | Contracts heading to recompete, with the incumbent identified |
| Closing-price ranges | Historical price ranges by agency and NAICS, for bid calibration |
| Forecast scorecard | Accuracy tracking of the platform's own forecasts, disclosed openly |
Installation¶
claude mcp add --transport http awardcast https://awardcast.ai/mcp
No account or key required. The surface is read-only public data.
Configuration¶
{
"mcpServers": {
"awardcast": {
"type": "http",
"url": "https://awardcast.ai/mcp"
}
}
}
Business Relevance¶
- Government contractors build bid pipelines from open solicitations and recompete flags without a dedicated research analyst
- Bid teams calibrate pricing against historical closing ranges by agency and NAICS
- Business development leads get named incumbents on every recompete target, which shapes capture strategy
- Agencies and primes research contractor award histories for past performance and teaming decisions
- Analysts track forecast accuracy against outcomes because the scorecard is published with the data
Integration with CorpusIQ¶
CorpusIQ's QuickBooks and Stripe connectors answer the money side of a government services business: invoicing, payment timing, and receivables. AwardCast answers the pipeline side those connectors cannot: which contracts exist, who holds them, and what they are worth. A GovCon operator runs AwardCast for opportunity discovery and CorpusIQ for the financial reconciliation of the work they win, in the same agent session.
Limitations¶
- Brand new listing (Aug 17, 2026), no track record yet
- Public contracting data only: no classified or restricted procurements
- Covers US federal awards; no state, local, or international contracting data
- No auth means no personalized pipeline state: pair with a CRM for tracking
- Hosted-only: no self-host option published