LeadHunting AI: Autonomous Real Estate Lead Generation for $0
The Problem
Real estate agents spend hours every day on lead generation — cold calling, following up, writing marketing content, posting on social media. It is repetitive, time-consuming, and takes them away from what they do best: closing deals.
I wanted to build a platform that handles all of this autonomously, using only free-tier APIs, so that any agent could afford it.
The Solution: LeadHunting AI
LeadHunting AI is an autonomous 24/7 AI-powered real estate lead generation platform. Here is how it works:
1. Autonomous Content Pipeline
The system auto-generates blog posts, images, audio, and video using Google AI Studio (free tier):
- Text: Gemini Flash generates SEO-optimized blog posts about real estate topics
- Images: Imagen 3 creates cover images for each post
- Audio: Google TTS creates MP3 narrations
- Video: Veo 2 creates video clips (async — polled later)
- Social: Auto-posts to LinkedIn, Twitter, Facebook with policy compliance
All of this runs on free-tier APIs. Total cost: $0/month.
2. AI-Powered Lead Scoring
Every lead is scored on a 0-100 scale based on:
- Lead type (buyer, seller, investor, CMA, tenant)
- Phone number presence
- TCPA consent
- Time of day
- Source channel
High-scoring leads are auto-qualified; low-scoring leads are marked lost.
3. Credit Billing System
The platform uses a credit-based billing model:
- Admin uses the $0 free tier (Google AI Studio)
- Users pay credits for AI generation (500%+ margin)
- Credits are purchased via Stripe (dynamic pricing via
price_data) - Atomic credit deduction with auto-refund on failure
4. Compliance Engine
Built-in compliance for:
- CAN-SPAM: Unsubscribe link + physical address in every email
- TCPA: Consent tracking + audit trail
- GDPR: Data portability (JSON export) + right to erasure
- Platform Policies: Rate limits for 9 social platforms
Tech Stack
| Layer | Technology |
|---|---|
| Frontend | Next.js 16, React 18, TypeScript, Tailwind CSS |
| UI | shadcn/ui component library |
| Backend | Prisma ORM, Supabase PostgreSQL |
| AI | Google AI Studio (Gemini Flash, Imagen 3, Veo 2, TTS) |
| Payments | Stripe (dynamic pricing) |
| Deployment | Vercel (free Hobby plan) |
| Auth | JWT with session invalidation |
Key Architecture Decisions
1. SPA + Serverless
The app uses a SPA-like architecture with Next.js App Router. Marketing pages are server-rendered for SEO; the dashboard is client-rendered for interactivity.
2. Built-in Cron Scheduler
No external cron service. A heartbeat endpoint (/api/cron/heartbeat) is triggered by real user traffic. Every page visit runs the scheduler, which fires overdue tasks (content pipeline, social posting, email sequences, etc.).
3. Fire-and-Forget with Keepalive
Background tasks use fetch with keepalive: true to trigger long-running operations (like the content pipeline) without blocking the heartbeat response. The pipeline runs in its own 60-second serverless function.
4. Silent Islamic Values Filter
All image generation uses a sanitizePromptForImageGeneration() function that enforces modest attire and male subjects only — without notifying users. This is a silent filter that aligns with the founder'''s values.
Results
- Traffic Goal: 1 million visitors/month (driven by autonomous content pipeline)
- Content Output: 4 blog posts/day (every 6 hours)
- Cost: $0/month (all free-tier APIs)
- Revenue: Credit billing at 500%+ margin
Try It
Visit leadhunting.app to start a 7-day free trial. No credit card required.
- 🌐 Website
- 📝 Blog
- 💰 Pricing
- 📖 Setup Guide
- 🐙 GitHub
This article was originally published on leadhunting.app.
Discussion
Have you built autonomous content pipelines before? What free-tier APIs do you use? Let me know in the comments!
Top comments (0)