subsidy4u

Atypis/subsidy4u

3.1

If you are the rightful owner of subsidy4u and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

Subsidy4U is an AI-powered assistant designed to help German companies and startups navigate over 2,000 subsidy programs by leveraging a model context protocol server.

Tools
5
Resources
0
Prompts
0

Subsidy4U - German Fördermittel Assistant

AI-powered assistant to help German companies and startups navigate 2,000+ subsidy programs (Fördermittel).

🎯 Problem

German companies struggle with the overwhelming landscape of subsidies:

  • 2,000+ programs across federal, state, and EU levels
  • Complex eligibility requirements
  • Fragmented information sources
  • Time-consuming manual research

💡 Solution

An AI assistant that:

  1. Extracts company details from websites or documents (HR-Auszug)
  2. Filters programs through conversational guidance
  3. Applies intelligent matching against eligibility criteria
  4. Visualizes the filtering process in real-time
  5. Delivers a curated list of relevant, likely-eligible programs

🏗️ Architecture

Split-Screen Interface

  • Left: Conversational AI assistant (AI SDK v5)
  • Right: Live visualization of filtering process

Tech Stack

  • Frontend: Next.js 15 + AI SDK v5 + TailwindCSS + shadcn/ui
  • Backend: Supabase (PostgreSQL + pgvector)
  • Deployment: Vercel + Supabase
  • Data Source: foerderdatenbank.de (scraped)

📦 Project Structure

subsidy4u/
├── browser-automation-mcp/    # Browser automation MCP server
│   ├── src/                   # TypeScript source
│   ├── dist/                  # Built JS files
│   └── docs/                  # Documentation
├── plan.md                    # Complete project plan
├── browser-mcp.md             # MCP server specification
└── NEXT_STEPS.md             # Implementation roadmap

🔧 Browser Automation MCP Server

v1.0.2 - Production-ready browser automation for AI assistants

Features

  • 24 browser automation tools powered by Playwright
  • Execute JavaScript for targeted data extraction
  • Network inspection to discover hidden APIs
  • Auto-pagination for batch scraping
  • Response size limits to prevent crashes
  • Comprehensive error handling

Quick Start

cd browser-automation-mcp
npm install
npx playwright install chromium
npm run build

Configure in Claude Desktop:

{
  "mcpServers": {
    "browser-automation": {
      "command": "node",
      "args": ["/path/to/subsidy4u/browser-automation-mcp/dist/index.js"]
    }
  }
}

See for full documentation.

📋 Implementation Status

  • ✅ Project planning - Complete architecture and roadmap
  • ✅ Browser MCP server - Built and tested
  • ⏳ Website scraping - Next phase
  • ⏳ Database setup - Supabase schema
  • ⏳ Frontend - Next.js + AI SDK v5
  • ⏳ Deployment - Vercel hosting

🚀 Next Steps

  1. Scrape foerderdatenbank.de

    • Use browser MCP to discover structure
    • Extract all ~2,000 programs with metadata
    • Parse legal requirements and filter attributes

  2. Build database

    • Design Supabase schema
    • Import scraped data
    • Add vector search for semantic matching

  3. Develop frontend

    • Implement chat interface
    • Build live filtering visualization
    • Create company info extraction flow

  4. Deploy

    • Launch on Vercel
    • Set up CI/CD
    • Monitor and iterate

📖 Documentation

  • - Complete project plan with architecture
  • - MCP server specification
  • - Implementation roadmap
  • - Scraping guide

🤝 Contributing

This is a personal project, but feedback and suggestions are welcome!

📄 License

MIT

Current Version: Browser MCP v1.0.2
Status: Development - MCP Server Complete
Repository: https://github.com/Atypis/subsidy4u