LeadForge is a full-stack, AI-powered SaaS platform that helps sales teams discover, analyze, and prioritize restaurant leads.
It combines data ingestion, enrichment, deterministic scoring, and AI-driven reasoning into a single intelligent system.
LeadForge is built as a hybrid microservices system:
- Google Places API ingestion
- Lead enrichment pipeline
- Deterministic scoring engine
- MongoDB database
- REST APIs
- Built with FastAPI + LangGraph
- Gemini-powered multi-agent reasoning
- AI scoring refinement
- Outreach message generation
- React + Vite + Tailwind
- Redux Toolkit + RTK Query
- SaaS dashboard for sales teams
LeadForge/
│
├── backend/ # Node.js API (data pipeline + scoring)
├── ai_service/ # Python FastAPI (AI reasoning layer)
├── frontend/ # React dashboard (UI + RTK Query)
├── RUNBOOK.md # Full setup & workflow guide
└── README.md
cd backend
npm install
npm run mongo:up
npm run devuv sync --project ai_service
uv run --project ai_service uvicorn ai_service.main:app --reload --port 8001cd frontend
npm install
npm run devGET /leadsPOST /scrapePOST /enrichPOST /score
POST /analyze-lead
- 🔍 Automated restaurant lead scraping (Google Places)
- 🧩 Data enrichment (website, social, sentiment)
- 📊 Deterministic lead scoring (0–100)
- 🤖 AI-powered reasoning (LangGraph + Gemini)
- 💬 Personalized outreach generation
- 📈 SaaS dashboard with filtering & pipeline tracking
Backend
- Node.js, Express, MongoDB
AI Layer
- FastAPI, LangGraph, LangChain
- Gemini API
Frontend
- React, Vite, Tailwind CSS
- Redux Toolkit, RTK Query
For complete setup instructions and service orchestration:
👉 See RUNBOOK.md
LeadForge is designed to function as a sales intelligence system that helps teams:
- Identify high-quality leads
- Understand business pain points
- Generate targeted outreach
- Close deals faster
Built as a real-world production system integrating backend engineering, AI orchestration, and frontend SaaS design.