SC0%
Back to Index
project

LeadGen

Desktop lead generation tool that scrapes 7+ sources using headless Chromium, scores businesses by online presence, and generates AI pitch angles via a local Ollama model.

Next.js 15Python 3.12Electron 36PlaywrightFastAPIConvexClerkOllama

LeadGen: Desktop Lead Generation OS

LeadGen is a comprehensive desktop application designed to streamline the process of finding and qualifying business leads at scale. By orchestrating multiple scrapers and local AI models, it provides a powerful, privacy-focused alternative to cloud-based lead generation services.

"LeadGen turns web scrapers into a fully automated sales pipeline, all while keeping your data local and secure."

Multi-Source Scraper Architecture

The heart of LeadGen is its robust scraping engine, which can target 7 distinct sources simultaneously through a round-robin proxy rotation system.

  1. Google Maps & Yelp: Deep business discovery and local SEO scoring.
  2. Yellow Pages, Manta, & Bing Places: Broad-spectrum directory coverage.
  3. Angi: Specialized lead discovery for home services and trades.
  4. Twitch: Targeted outreach for streamers and specialty niches.

Local AI & Ollama Integration

One of LeadGen's unique features is its integration with Ollama for local AI processing. It uses the qwen2.5 model to analyze lead data and generate personalized pitch angles.

  • Zero Data Leakage: All analysis happens on-device, ensuring client privacy.
  • Dynamic Pitch Generation: Tailors outreach based on the specific business niche and online presence scores.
  • In-App Model Management: One-click installer and model downloader for Windows users.

Credit & Subscription System

LeadGen implements a sophisticated "hold and settle" credit model, allowing for transparent usage tracking and auto-refunds on failed scans.

  • Tiered Access: Supports free, starter, pro, and premium subscription tiers.
  • Automated Provisioning: Integrated with Cash App via a custom Google Apps Script email parser.
  • Rate Limiting: Protects both the user and the scrapers through intelligent request throttling.

Technical Stack

  • Frontend: Next.js 15, React 19, Tailwind CSS v4.
  • Backend: Python 3.12, FastAPI, Playwright (Headless Chromium).
  • Desktop: Electron 36 with bundled Python server.
  • Database: Convex (Cloud Sync) & SQLite (Local Lead Storage).
  • Auth: Clerk with JWT integration.
  • AI: Ollama (Local LLM Orchestration).

Project Roadmap

LeadGen is currently fully functional, with future plans for deeper CRM integrations and expanded AI reasoning capabilities for automated outreach campaigns.

https://github.com/bibbisalsd/leadgen