Awardopedia - Federal Contracts
LocalAbout Awardopedia - Federal Contracts
Awardopedia is a web interface for searching U.S. federal contract award data from USASpending.gov. It also offers AI contract analysis using Claude, with analysis paid for by credits.
Key features
- 01Search and filter contract awards
- 02Browse USASpending.gov data
- 03Claude-based contract analysis
- 04Weekly data synchronization
Use cases
- 01Find federal contract award records
- 02Review contract summaries and insights
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
npx -y awardopedia-mcpAWARDOPEDIA_API_KEYSecretFree API key from awardopedia.com/signupLocal servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
2Saved tool list · updated 2w ago
| Parameter | Type | Required | Description |
|---|---|---|---|
| q | string | No | Search keywords (e.g., 'cybersecurity', 'janitorial services', 'IT support') |
| limit | number | No | Number of results (1-25, default 10) |
| naics | string | No | NAICS code filter (e.g., '541512' for IT services, '561720' for janitorial) |
| state | string | No | Two-letter state code for place of performance (e.g., 'VA', 'CA', 'TX') |
| set_aside | string | No | Set-aside type filter (e.g., 'SBA', '8(a)', 'SDVOSB', 'WOSB', 'HUBZone') |
Details
From the project READMEAwardopedia
Free, searchable interface for U.S. federal contract award data sourced from USASpending.gov.
Tech Stack
- Frontend: React 18 + Vite 5, react-router-dom v6, Chart.js, lucide-react
- Backend: Node.js 20 + Express 4, PostgreSQL 15
- AI: Anthropic Claude API for contract analysis (paid credits)
- Payments: Stripe Checkout for credit packs
- Auth: Magic link email via SendGrid + JWT
- Hosting: DigitalOcean App Platform
- DNS/CDN: Cloudflare
Local Development
Prerequisites
- Node.js 20+
- PostgreSQL 15+
- Python 3.10+ (for data sync scripts)
Setup
# Clone
git clone https://github.com/openclaw/awardopedia.git
cd awardopedia
# Database
psql -d your_db -f scripts/schema.sql
# API
cd api
cp ../.env.example .env # Edit with your values
npm install
npm run dev # Starts on :3001
# Frontend (separate terminal)
cd web
npm install
npm run dev # Starts on :3000, proxies /api to :3001
Environment Variables
See .env.example for all required variables. At minimum for local dev:
DATABASE_URL=postgresql://localhost:5432/awardopedia
JWT_SECRET=any-dev-secret
PORT=3001
Project Structure
awardopedia/
├── api/ # Express API server
│ ├── src/
│ │ ├── db/ # Database connection + queries
│ │ ├── middleware/ # Auth, credits, rate limiting
│ │ ├── routes/ # API route handlers
│ │ └── services/ # Claude AI, Stripe, SendGrid
│ └── Dockerfile
├── web/ # React + Vite frontend
│ ├── src/
│ │ ├── components/ # Reusable UI components
│ │ ├── pages/ # Route pages
│ │ ├── styles/ # CSS tokens, global, components
│ │ └── utils/ # API client, formatters, SEO
│ └── public/
├── scripts/
│ ├── schema.sql # PostgreSQL schema
│ ├── ingest.py # Bulk CSV ingestion
│ └── sync.py # Weekly USASpending API sync
├── docs/
│ ├── api.md # API endpoint reference
│ ├── data-model.md # Database schema docs
│ └── deployment.md # Deployment guide
└── assets/ # Brand logos
Data
All contract data is sourced from USASpending.gov, the official open data source of federal spending information maintained by the U.S. Department of the Treasury as mandated by FFATA.
- Bulk import:
python3 scripts/ingest.py --file contracts.csv - Weekly sync:
python3 scripts/sync.py(fetches last 7 days from USASpending API)
Business Model
- Core data access (search, filter, browse): Free
- AI analysis (contract insights, summaries): Paid credits
- Starter: 100 credits / $9
- Pro: 500 credits / $29
- Power: 2,000 credits / $79
License
Proprietary. Data sourced from USASpending.gov (public domain).
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.