Automated German startup discovery with AI-powered scoring - Find newly registered tech companies from the German commercial register (Handelsregister) with intelligent relevance scoring.
Automatically scans the German Handelsregister (commercial register) for newly registered tech startups across 10 major cities, scores them by relevance (0-100), and delivers actionable leads for B2B sales, investment, or partnership opportunities.
Perfect for: B2B SaaS companies, angel investors, VCs, business consultants, market researchers.
# Health check
curl https://startup-radar-bot-test.vercel.app/api/mock
# Test scan
curl -X POST https://startup-radar-bot-test.vercel.app/api/mock \
-H "Authorization: Bearer demo-key" \
-H "Content-Type: application/json" \
-d '{"days_back":30,"max_results":5,"min_score":40}'🌐 Interactive Demo: https://startup-radar-bot-test.vercel.app
💡 The demo uses sample data (15 realistic German startups). The production version connects to the real Handelsregister.ai API.
- Automated scanning of German commercial register via Handelsregister.ai
- 10 major German cities: Berlin, München, Hamburg, Frankfurt, Darmstadt, Köln, Stuttgart, Düsseldorf, Leipzig, Karlsruhe
- Rotating coverage: 5 cities per run (Group A/B weekly rotation)
- Filters by legal form (GmbH, UG, AG) and tech keywords
Each startup receives a relevance score based on:
- Company Age (30 pts) - Newer = Higher score
- Tech Keywords (25 pts) - Software, AI, SaaS, FinTech, etc.
- Regional Priority (20 pts) - Major tech hubs prioritized
- Legal Form (10 pts) - GmbH/UG preferred
- Description Quality (15 pts)
Priority Levels:
- 🔴 High (≥70) - Immediate action
- 🟡 Medium (≥40) - Worth reviewing
- 🟢 Low (<40) - Optional follow-up
- Make.com workflow orchestration
- Bi-weekly automatic scans (Monday + Thursday)
- Google Sheets data export
- Claude AI integration for pain point analysis
- Zero manual intervention
- ~€2/month operational costs
- Free tiers: Vercel, Make.com (1000 ops), Handelsregister.ai (500 credits)
- Serverless architecture - no infrastructure management
| Component | Technology |
|---|---|
| Backend | Python 3.9+ |
| Hosting | Vercel (Serverless) |
| Data Source | Handelsregister.ai API |
| Automation | Make.com |
| AI Analysis | Anthropic Claude 4.5 |
| Storage | Google Sheets |
Or manually:
# Clone repository
git clone https://github.com/Mutigen/startup-radar-bot-test.git
cd startup-radar-bot-test
# Deploy with Vercel CLI
npm i -g vercel
vercel --prodFor the production API (/api/scan), add these in Vercel:
# API Key for Make.com authentication
API_KEY=your_secure_random_api_key
# Handelsregister.ai API Key
HANDELSREGISTER_API_KEY=your_handelsregister_api_key💡 Mock API (
/api/mock) doesn't need environment variables!
# Health check
curl https://your-deployment.vercel.app/api/mock
# Test scan
curl -X POST https://your-deployment.vercel.app/api/mock \
-H "Authorization: Bearer demo-key" \
-H "Content-Type: application/json" \
-d '{"days_back":60,"max_results":10,"min_score":40}'https://your-deployment.vercel.app
Health check - returns service status.
Response:
{
"status": "online",
"service": "Startup Radar Bot (MOCK DEMO)",
"version": "2.2.0 (Mock)",
"current_group": "B",
"note": "This is a DEMO API with sample data."
}Scan for startups (demo data).
Request:
{
"days_back": 90, // Days to look back (default: 30)
"max_results": 10, // Max startups (default: 30)
"min_score": 40 // Min score (default: 20)
}Response:
{
"success": true,
"count": 5,
"results": [
{
"startup_name": "TechVision AI GmbH",
"city": "Berlin",
"relevance_score": 85,
"tags": "Berlin,Software,KI/AI",
"founded_year": "2025",
"short_description": "AI-powered enterprise software...",
"handelsregister_id": "HRB234567"
}
]
}Full schema: See API Documentation
Automated weekly delivery of qualified tech startup leads with AI-analyzed pain points.
Bi-weekly scans with relevance scoring for quick filtering of early-stage investment opportunities.
Historical data collection for startup ecosystem trend analysis.
Identify potential technology partners or acquisition targets by keywords and business description.
Make.com Scheduler
↓ (Monday + Thursday, 8:00 AM)
Vercel Serverless API (/api/scan)
↓ (Search 5 cities)
Handelsregister.ai API
↓ (Calculate scores, filter)
JSON Response (top 30)
↓ (Iterator)
Make.com → Claude AI Analysis
↓ (Pain points, insights)
Google Sheets (CRM)
City Rotation:
- Even weeks (2, 4, 6...): Group A (Berlin, München, Hamburg, Frankfurt, Darmstadt)
- Odd weeks (1, 3, 5...): Group B (Köln, Stuttgart, Düsseldorf, Leipzig, Karlsruhe)
Mock API returns 15 realistic startups:
| City | Example Startups | Industries |
|---|---|---|
| Berlin | TechVision AI, RoboTech | AI, Robotics |
| München | CloudFlow, GreenEnergy | SaaS, GreenTech |
| Hamburg | DataMind, FoodTech | Analytics, FoodTech |
| Frankfurt | FinTech Innovations, InsurTech | FinTech, Insurance |
| Darmstadt | SmartLogistics, SmartCity | Logistics, IoT |
Score Range: 40-90 points
Registration Dates: Oct 2024 - Jan 2025
# Start mock API locally
python api/mock_scan.py
# Test locally
curl -X POST http://localhost:8000/api/scan \
-H "Authorization: Bearer test-key" \
-H "Content-Type: application/json" \
-d '{"days_back":30,"max_results":5,"min_score":40}'# Run test suite
python test_mock_api.py
# Test deployed API
./test_deployment.sh https://your-deployment.vercel.app- Testing Guide - Complete testing documentation
- Quick Start - 10-minute setup guide
- API Reference - Endpoint documentation
Contributions welcome! Please:
- Fork the repository
- Create a feature branch (
git checkout -b feature/AmazingFeature) - Commit changes (
git commit -m 'Add AmazingFeature') - Push to branch (
git push origin feature/AmazingFeature) - Open a Pull Request
Ideas:
- Add more regions (Austria, Switzerland)
- Email enrichment integration
- LinkedIn profile scraping
- Notion database sync
- Web dashboard with analytics
- v1.0 - Core functionality (Nov 2025)
- v2.0 - Multi-city expansion (10 cities)
- v2.2 - City rotation system (Jan 2025)
- v2.3 - Email enrichment (Hunter.io/Apollo)
- v2.4 - LinkedIn automation (Phantombuster)
- v3.0 - Web dashboard with analytics
- v3.1 - Expand to Austria & Switzerland
| Service | Free Tier | Monthly Cost |
|---|---|---|
| Vercel | Unlimited deployments | €0 |
| Handelsregister.ai | 500 credits/month | €0-2 |
| Make.com | 1000 operations | €0 |
| Google Sheets | Unlimited | €0 |
| Total | ~€0-2/month |
For typical usage: 2 scans/week, 30 results per scan
This project is licensed under the MIT License - see the LICENSE file for details.
- Handelsregister.ai - German company registry API
- Anthropic - Claude AI for analysis
- Make.com - Workflow automation
- Vercel - Serverless hosting
- Issues: GitHub Issues
- Live Demo: https://startup-radar-bot-test.vercel.app
Built with ❤️ by MAMIKO @ MUT-i-GEN
Automated startup scouting for smart entrepreneurs, investors, and B2B companies.
⭐ Star this repo if you find it useful!