AI agent for discovering micro-product opportunities using the "existing demand" methodology: find products with proven users → extract ONE function → make it simpler.
- 🔍 Searches multiple sources (Chrome Web Store, Product Hunt) across specified niches
- 📊 Finds products with proven demand (10k+ users, good ratings)
- 🔬 Deep analysis of each product:
- Features, target audience, pricing model
- Review mining: extracts recurring complaints, missing features, UX issues
- Identifies weaknesses and simple use cases not handled well
- 💡 Generates micro-product ideas in "one function" format:
- Concrete input→output examples
- Evidence of existing demand
- How it differs from bloated competitors
- 📈 Scores each idea (1-10) on:
- Existing demand, time-saving potential, monetization
- Competition saturation, MVP simplicity, personal fit
- 📝 Outputs sorted by score:
- JSON for programmatic use
- Markdown summary for human review
# Install dependencies
npm install
# Run in mock mode (no API keys required)
npm run mine
# Check results
cat out/ideas-summary-*.md# Single niche
npm run mine -- --niche="YouTube creators"
# Multiple niches (semicolon-separated)
npm run mine -- --niches="YouTube creators;Developers and programmers"
# Custom search query
npm run mine -- --query="AI writing assistant"
# Set minimum score threshold (default: 5)
npm run mine -- --min-score=7
# Enable Google Trends analysis (enabled by default in real mode)
npm run mine -- --enable-trends
# Skip Google Trends analysis
npm run mine -- --skip-trends
# Filter by minimum search demand (requires --enable-trends)
npm run mine -- --enable-trends --min-search-demand=6- Setup - full installation guide
- Usage - examples and CLI options
- Configuration - all environment variables
- Architecture - code structure and extension
Each idea is scored on 8 dimensions (1-10 scale):
| Metric | Weight | Description |
|---|---|---|
| Existing Demand | 18% | Competitors, user counts |
| Time Saving | 12% | How much time does user save |
| Monetization | 18% | B2B/creator audience, willingness to pay |
| Competition | -10% | How crowded (high = bad) |
| Simplicity | 13% | Can build MVP in weekend-month |
| Personal Fit | 9% | Matches your skills |
| Platform Risk | -15% | ToS/policy risk (high = bad) |
| Search Demand | 15% | Google Trends data for US/Europe |
finalScore = weighted sum, normalized to 1-10 scale.
Ideas with platformRisk >= 7 or matching risky patterns (downloaders, mass actions) are auto-filtered.
When --enable-trends is active, each idea gets:
- Primary keyword: main search query users would type
- Secondary keywords: related long-tail queries
- Suggested name: SEO-friendly product name
- Search regions: US + major EU countries (GB, DE, FR, ES, IT, NL, SE, PL)
This data helps validate search demand and plan SEO strategy before building.
Note: Google Trends API has strict rate limits. If you see many failures, the system will automatically use fallback scores (searchDemand=4) and continue processing. Keyword generation uses
OPENAI_API_KEY(same as browser automation).
- 🎬 YouTube creators
- 📱 Instagram Reels editors
- 🎵 TikTok content creators
- 📊 Excel/Google Sheets power users
- ✍️ Copywriters and content marketers
- 🛒 Small e-commerce owners
- 💻 Developers and programmers
- 🎓 Students and researchers
- ✅ Chrome Web Store (browser extensions)
- 🔜 Product Hunt (coming soon)
MIT