The MCP-native federal & SLED intelligence platform
SCOUT plugs federal and SLED market intelligence directly into Claude, ChatGPT, and any MCP client. MCP isn't a feature we added — it's how SCOUT is built.
What is the SCOUT MCP server?
SCOUT's MCP server lets any MCP-enabled AI assistant query SCOUT's live federal and SLED market intelligence in natural language.
Once connected, your AI can run SCOUT's full toolset — opportunity and solicitation search, recompete pipeline tracking, award-history and competitive analysis, and teaming match — over a dataset of 100K+ opportunities across 50+ federal agencies, 560K+ active contract awards, and 2.5M+ unique awardees. SCOUT reads the actual PWS and SOW attachments, expands the acronyms keyword search misses, and scores relevance to your firm before returning results, so the model works from decision-ready intelligence instead of raw records it has to interpret.
Why MCP-native matters
SCOUT was architected MCP-first. The model talks to SCOUT's intelligence layer directly — it doesn't scrape exports or wrap a third-party data broker.
While others connected a server to an existing database, SCOUT was built on MCP from the ground up — not bolted on. That means document intelligence, vocabulary-aware search, incumbent identification, and relevance scoring all run inside the protocol, so what reaches your AI is finished analysis rather than a data dump to summarize.
Built MCP-first
The protocol is the core architecture, not an afterthought added to a legacy platform.
Direct to intelligence
The model queries SCOUT's scoring and classification layer — not a raw record feed.
Not a broker wrapper
No third-party data middleman between your AI and SCOUT's federal + SLED coverage.
Ask SCOUT in plain language
Every SCOUT capability is a callable tool inside your AI assistant. Ask the way your capture team actually thinks.
Recompete pipeline
Award-history analysis
Opportunity discovery
Teaming match
Protest intelligence
Market sizing
Works with the AI clients your team already uses
Plus custom agents and any client that supports Streamable HTTP or SSE. Consumer Google Gemini and Microsoft Copilot don't yet support remote MCP servers.
The dataset behind every answer
100K+
Opportunities
50+
Federal Agencies
560K+
Active Awards
2.5M+
Unique Awardees
98%+
of Federal Dollars (≥ $100K floor)
Real-time
Updates
Data floor of ≥ $100K captures 98%+ of federal contract dollars. Federal and SLED coverage, refreshed continuously.
Three steps to SCOUT inside your AI
No engineering ticket. Connect in minutes and start asking.
Get an account & API key
Sign up at scout.primerfp.com on a tier with MCP enabled, then create an API key in Account Settings → API Keys — it starts with prfp_.
Add the SCOUT MCP endpoint
Drop the endpoint into your client's MCP config with an Authorization Bearer header (snippet below), then restart the client.
Ask SCOUT
Start a conversation and ask “What GovCon tools do you have?” — then put your pipeline to work.
{
"mcpServers": {
"scout": {
"type": "http",
"url": "https://mcp.primerfp.com/mcp",
"headers": {
"Authorization": "Bearer prfp_YOUR_KEY_HERE"
}
}
}
}Streamable HTTP endpoint: https://mcp.primerfp.com/mcp · SSE fallback: https://mcp.primerfp.com/sse. Claude Desktop connects via the mcp-remote bridge — see the setup guide.
MCP access starts at the MCP Explorer tier of SCOUT MCP ($29/mo). A free MCP Trial is available, and Tactics, Strategic, and Premier include MCP with higher limits.
See pricingSCOUT MCP, answered
Does SCOUT have an MCP server?
Yes. SCOUT runs a production MCP server that lets any MCP-enabled AI assistant query SCOUT's live federal and SLED market intelligence in natural language. Connect it to Claude, ChatGPT, Cursor, or any MCP-compatible client and ask questions about opportunities, recompetes, awards, competitors, and teaming partners.
What is an MCP-native intelligence platform?
MCP-native means SCOUT was architected on the Model Context Protocol from the ground up, rather than having a connector bolted onto a legacy database. The model talks directly to SCOUT's intelligence layer — which reads solicitation attachments, expands GovCon acronyms, and scores relevance to your firm — instead of receiving raw exports it has to interpret. MCP isn't a feature we added; it's how SCOUT is built.
Which AI clients work with SCOUT's MCP?
SCOUT works with Claude (Desktop via mcp-remote, and Claude Code), ChatGPT Desktop, Cursor, Windsurf, VS Code (Copilot), n8n, and any client that supports Streamable HTTP or SSE. Consumer Google Gemini and consumer Microsoft Copilot do not yet support remote third-party MCP servers; ChatGPT Desktop and Cursor are the easiest on-ramps today.
Does SCOUT cover SLED as well as federal?
Yes. SCOUT covers both federal procurement (SAM.gov, USASpending, FPDS-derived award data) and state, local, and education (SLED) opportunity data — all queryable through the same MCP server. You can ask about federal recompetes and SLED solicitations in the same conversation.
How is SCOUT's MCP different from a bolted-on MCP connector?
A bolted-on connector wraps an existing platform or a third-party data broker and pipes raw records to the model. SCOUT was architected MCP-first, so the model queries SCOUT's intelligence layer directly — document-level classification, vocabulary-aware search, incumbent identification, and relevance scoring all happen before results are returned. You get decision-ready intelligence, not raw JSON the model has to summarize.
How much does SCOUT MCP access cost?
MCP access starts at the MCP Explorer tier of SCOUT MCP ($29/mo), which puts SCOUT intelligence inside ChatGPT, Claude, and other MCP-ready clients without the full web portal. A free MCP Trial is available, and the Tactics, Strategic, and Premier tiers include MCP access with higher limits and more tools. The web-only Free plan does not include MCP. See the pricing page for current tiers.
How do I connect SCOUT to Claude or ChatGPT?
Create a SCOUT account with MCP enabled, generate an API key in Account Settings (it starts with prfp_), then add the SCOUT MCP endpoint (https://mcp.primerfp.com/mcp) to your client's config with an Authorization: Bearer header. Restart the client and ask it what GovCon tools it has. Full per-client setup is in the SCOUT MCP documentation.
Put SCOUT intelligence inside your AI.
MCP-native federal & SLED intelligence. Free trial, no engineering required, connected in minutes.