Lighthouse AI is a high-performance marketing and analysis engine designed for the modern real estate landscape. By bridging the gap between creative storytelling and technical investment logic, Lighthouse transforms a simple property URL into a complete, cross-platform marketing kit in seconds.
🚀 Live Demo: here
The engine utilizes a dual-persona architecture to serve different market needs:
- Ivy Aria (Professional): Crafts luxury narratives and lifestyle-driven copy. Focused on emotional resonance and "vibe" for residential buyers.
- Ryan Alexander (Executive): Delivers data-heavy, technical blueprints. Focused on ROI, market positioning, and investment logic.
- Framework: Next.js 16 (App Router) with Turbopack.
- AI Orchestration: Vercel AI SDK.
- Inference: GPT-OSS 120B running on Groq LPUs (Low-latency, high-throughput).
- Data Extraction: Firecrawl (Autonomous web scraping and markdown conversion).
- Validation: Zod-based schema enforcement for consistent API responses.
- Styling: Tailwind CSS with a "Glassmorphic" Blueprint aesthetic.
- Autonomous Scraping: Enter a Zillow, Realtor.com, or Redfin URL to automatically ingest property specs and descriptions.
- Multi-Channel Output: Generates official MLS drafts, Instagram captions, and TikTok storyboards simultaneously.
- Feature Chip Extraction: Intelligently identifies top-selling points (e.g., "Quartz Countertops", "New HVAC") for quick-glance UI.
- Adaptive UI: A "Lighthouse" blueprint background with a glassmorphic interface that responds to the selected persona.
git clone https://github.com/codelikeagirl29/lighthouse.git
cd lighthouseCreate a .env.local file in the root:
OPENAI_API_KEY=your_openai_key
GROQ_API_KEY=your_groq_key
FIRECRAWL_API_KEY=your_firecrawl_keynpm install
npm run dev- Virtual Staging Integration: Connecting the Shadow + Slate canvas engine for automated floorplan visualization.
- Direct Social API: One-click posting to Instagram and Facebook Business suites.
- PDF Export: Generation of professional "Listing Flyers" for open house events.
Distributed under the MIT License. See LICENSE for more information.
**Developed by Lindsey Howard