GrantIQ MCP

GrantIQ MCP

A paid MCP server that gives AI agents instant, structured access to every open US federal grant and contract opportunity.

Category
Visit Server

README

GrantIQ MCP

A paid MCP server that gives AI agents instant, structured access to every open US federal grant and contract opportunity. Backed by Grants.gov + SAM.gov, with 24-hour SQLite caching, per-key daily limits, and a Stripe-friendly upgrade hook.

https://mcp-grantiq.up.railway.app/mcp


MCPize listing copy

Title: GrantIQ MCP — US Federal Grants & Contracts for AI Agents

Subtitle: Search, match, and monitor every open SAM.gov and Grants.gov opportunity from any LLM, in one tool call.

GrantIQ MCP turns the messy world of US government funding into a clean, agent-ready API. Eight typed tools cover everything from broad keyword search and NAICS-coded contract discovery to deadline radar and historical award lookups. Responses are normalized across both data sources, returned as plain JSON, and cached for 24 hours so repeat queries don't burn rate limit on the upstream agencies.

Built for procurement copilots, grant-finder agents, and B2G sales tools, GrantIQ MCP ships with a free tier (50 calls/day) that any developer can drop into Claude or Cursor in 15 seconds. Upgrade to Pro for $19/mo for unlimited calls, or pay $0.02 per call for bursty workloads — billing flips a key from free to pro via a single webhook, with no app rewrites required.


Pricing

Tier Price Limit
Free $0 50 calls / day / key
Pro $19 / month Unlimited
Pay-as-you-go $0.02 / call Unlimited

Install

claude mcp add grantiq-mcp --url https://mcp-grantiq.up.railway.app/mcp

Then add the API key Claude should use:

claude mcp env grantiq-mcp X-API-Key=grantiq-dev-key-001

The dev key grantiq-dev-key-001 is seeded automatically and capped at 50 calls/day. Mint your own with the admin endpoint below.


Tool reference

Tool Args Returns
search_grants keyword, agency, amount_min, amount_max, status, limit List of matching grants
get_grant_details opportunity_id Full grant object
search_contracts keyword, naics_code, agency, set_aside, limit List of SAM contracts
get_contract_details notice_id Full contract object
match_opportunities org_description, org_type, focus_areas Top 10 matched opps (grants + contracts)
get_agencies type (grants / contracts / both) Agencies with open opps, ranked by volume
get_deadlines days_ahead Opportunities closing within N days
search_awards recipient_name, agency, year Past SAM award history

Every tool returns either:

{ "success": true,  "cached": false, "data": { ... } }

or, on failure:

{ "success": false, "error": "..." }

When the daily limit is hit, the HTTP layer returns 429 with body:

{ "error": "Limit reached. Upgrade at mcpize.com/grantiq-mcp" }

curl examples

The MCP endpoint speaks the JSON-RPC streamable-HTTP protocol and is best used through an MCP client. For convenience, every tool also has a REST mirror at /tools/<tool>.

# Search grants
curl -s https://mcp-grantiq.up.railway.app/tools/search_grants \
  -H 'Content-Type: application/json' \
  -H 'X-API-Key: grantiq-dev-key-001' \
  -d '{"keyword":"renewable energy","limit":5}'

# Get grant details
curl -s https://mcp-grantiq.up.railway.app/tools/get_grant_details \
  -H 'Content-Type: application/json' \
  -H 'X-API-Key: grantiq-dev-key-001' \
  -d '{"opportunity_id":"345678"}'

# Search SAM.gov contracts (cyber, small-business set-aside)
curl -s https://mcp-grantiq.up.railway.app/tools/search_contracts \
  -H 'Content-Type: application/json' \
  -H 'X-API-Key: grantiq-dev-key-001' \
  -d '{"keyword":"cybersecurity","set_aside":"SBA","limit":5}'

# Get contract details
curl -s https://mcp-grantiq.up.railway.app/tools/get_contract_details \
  -H 'Content-Type: application/json' \
  -H 'X-API-Key: grantiq-dev-key-001' \
  -d '{"notice_id":"abc123"}'

# Match opportunities to an org profile
curl -s https://mcp-grantiq.up.railway.app/tools/match_opportunities \
  -H 'Content-Type: application/json' \
  -H 'X-API-Key: grantiq-dev-key-001' \
  -d '{"org_description":"Battery storage R&D startup","org_type":"for-profit","focus_areas":["energy","battery","storage"]}'

# Top agencies right now
curl -s https://mcp-grantiq.up.railway.app/tools/get_agencies \
  -H 'Content-Type: application/json' \
  -H 'X-API-Key: grantiq-dev-key-001' \
  -d '{"type":"both"}'

# Deadlines in the next 14 days
curl -s https://mcp-grantiq.up.railway.app/tools/get_deadlines \
  -H 'Content-Type: application/json' \
  -H 'X-API-Key: grantiq-dev-key-001' \
  -d '{"days_ahead":14}'

# Past award history
curl -s https://mcp-grantiq.up.railway.app/tools/search_awards \
  -H 'Content-Type: application/json' \
  -H 'X-API-Key: grantiq-dev-key-001' \
  -d '{"recipient_name":"Acme Corp","year":2025}'

Operator endpoints

Mint a fresh key (free or pro):

curl -s https://mcp-grantiq.up.railway.app/admin/keys \
  -H 'X-Admin-Token: $GRANTIQ_ADMIN_TOKEN' \
  -H 'Content-Type: application/json' \
  -d '{"tier":"free","daily_limit":50}'

Promote a free key to pro (called from the Stripe checkout success webhook):

curl -s https://mcp-grantiq.up.railway.app/billing/upgrade \
  -H 'X-Admin-Token: $GRANTIQ_ADMIN_TOKEN' \
  -H 'Content-Type: application/json' \
  -d '{"key":"grantiq_xxxxx"}'

Local dev

python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
python server.py        # http://localhost:8000
pytest -v               # 22 tests, hermetic with respx mocks

Deploy to Railway

The repo includes nixpacks.toml. After railway init, just railway up. The start command is python server.py (not uvicorn ...) because Railway exec's [start].cmd without a shell — server.py reads $PORT via os.getenv and starts uvicorn programmatically.

Set these on the Railway service:

  • SAM_API_KEY — your SAM.gov key
  • GRANTIQ_ADMIN_TOKEN — long random string
  • GRANTIQ_DB_PATH — /data/grantiq.db if you mount a Volume (recommended so cache + keys survive deploys)

A ready-to-run deploy.sh is in this directory.


Caveats

  • SAM.gov award history is exposed via the opportunities API filtered to ptype=a (Award Notice). For deeper award analytics (CFDA-level totals, recipient rollups), USAspending.gov is more appropriate and would be a v2 add.
  • Cache TTL is global (24h). For deadline-sensitive flows the agent should pass distinct query args or call /admin/keys to mint a key that bypasses cache (not yet exposed — see TODO).
  • Match scoring is deterministic Jaccard over title + agency tokens. It's fast and free; an LLM rerank can be added behind the same interface.
  • Stripe wiring is one-way (/billing/upgrade). The MCP server doesn't host the checkout page itself — that lives on the marketing site.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured