You have a startup idea. You think it's original. You spend weeks building a pitch deck, refining your value proposition, maybe even writing code. Then you Google it and find three YC-backed companies already doing the same thing, funded two years ago, with teams three times your size.
This is the startup validation gap: founders invest time and energy into ideas without visibility into the competitive landscape. By the time they discover overlap, they've already committed. The information exists, scattered across databases, pitch decks, and Crunchbase profiles, but no one is synthesizing it before you start building.
BeenDone closes that gap. It takes your startup idea, searches a vectorized database of YC-backed companies, and returns an AI-powered competitive analysis complete with feature comparisons, market positioning, and actionable pivot suggestions, all before you write a single line of code.
Enter your startup title and a short description on the landing page. BeenDone takes both inputs and prepares them for semantic search.
Your idea is embedded using a sentence-transformer model and compared against a MongoDB Atlas database of YC-backed startups using vector search. The top 10 most semantically similar companies are returned, ranked by cosine similarity.
The matched startups are passed to Google Gemini, which generates a structured evaluation:
- Competitor Breakdown: Direct competitors with strengths and weaknesses.
- Feature Metrics: Radar chart scoring your idea across AI technology, UX, social features, personalization, and analytics.
- Inline Suggestions: Specific text replacements for your description that add differentiation, matched to the exact phrasing you wrote.
- Pivot Suggestions: Strategic recommendations for repositioning your idea based on gaps in the existing market.
- Success Likelihood: A confidence score reflecting how saturated the space already is.
The analysis is rendered in a full-screen dashboard with:
- Similar Startups Panel: Scrollable list of matched companies with one-liners, stages, and similarity scores.
- Radar Chart: Visual comparison of your idea's feature coverage against the competitive average.
- Funding Distribution: Pie chart breaking down competitor funding across ranges.
- Stage Distribution: Pie chart showing the maturity breakdown of similar startups.
- Heat Map: Leaflet-powered geographic visualization of where competitors are headquartered.
- Industry Readiness Bar: A progress indicator for your idea's likelihood of success.
- AI Chatbot: A contextual assistant that can answer follow-up questions about your analysis.
Click "Generate Report" to pass your analysis to a second Gemini call that produces a structured venture fund proposal. The report is rendered as a downloadable PDF with executive summary, market opportunity, and growth strategy sections.
Enter Startup Idea
|
Vector Search Against YC Database
|
Gemini Evaluates Competitive Landscape
|
Dashboard Renders Analysis
|
|--- Competitor List
|--- Feature Radar Chart
|--- Funding & Stage Distribution
|--- Geographic Heat Map
|--- Inline Suggestions
|
AI Chatbot for Follow-Up Questions
|
Generate Venture Fund Proposal (PDF)
| Layer | Technology |
|---|---|
| Frontend | React / Next.js 15, TypeScript, Tailwind CSS |
| Backend | FastAPI, Pydantic, Uvicorn |
| Database | MongoDB Atlas (Vector Search) |
| Embeddings | Sentence-Transformers (all-MiniLM-L6-v2) |
| AI Model | Google Gemini 2.5 Pro |
| PDF Generation | @react-pdf/renderer |
| Maps | Leaflet.js |
| AI Chat | Vercel AI SDK + Google Generative AI |
- Python 3.9+
- Node.js 18+
Create a .env file in the root directory:
MONGODB_URI=your_mongodb_connection_string
DB_NAME=your_database_name
COLLECTION_NAME=your_collection_name
GEMINI_API_KEY=your_gemini_api_keycd back-end
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
pip install -r requirements.txt
python app.pyThe API runs at http://localhost:8000. Docs at http://localhost:8000/docs.
cd front-end
npm install
npm run devThe app runs at http://localhost:3000.