AI-powered entertainment discovery — find your next watch based on your mood.
lumigo.tv · AI Mood Search · Blog
Lumigo is an entertainment discovery platform that uses NLP to match movies, TV series, and anime to your mood. Instead of scrolling through endless catalogs, you describe how you feel or what you're in the mood for, and the AI agent finds the right content.
The problem: Every streaming platform optimizes for engagement, not for what you actually want to watch right now. Recommendation algorithms push trending content, not personal taste. The result: 30 minutes browsing, 0 minutes watching.
The solution: A mood-first search engine for entertainment. Type "something slow and melancholic, like Lost in Translation" and get results that match the vibe — not just the genre tags.
- AI Mood Search — Natural language queries matched to 10M+ titles using NLP. Describe a feeling, a scene, a vibe — the AI finds it.
- Curated lists — Create, share, and discover community-built collections.
- Ratings & tracking — Track watched, watching, and planned content across movies, series, and anime.
- SEO blog engine — AI-assisted content generation targeting long-tail entertainment queries. Hundreds of indexed posts driving organic traffic.
| Metric | Value |
|---|---|
| Unique users | 30,000+ |
| Page views | 500,000+ |
| Time to traction | 11 months |
| Marketing spend | €0 |
| Content database | 10M+ movies & TV shows |
All growth was organic — Reddit, SEO, and word of mouth. Zero ads, zero influencer deals, zero paid campaigns.
Lumigo is a closed-source product. This repository serves as a public reference for the project.
Stack highlights:
- Full-stack web application with server-side rendering
- NLP pipeline for mood-based semantic search
- AI content generation for blog and recommendations
- Integration with external movie/TV databases (10M+ titles)
- Multi-language support
I co-founded Lumigo with Lorenzo — we were both tired of spending more time browsing than watching. The AI mood search started as an experiment and became the core product.
The real learning wasn't the product — it was distribution. Getting 500K page views with €0 taught me how to find users where they already complain about the problem (Reddit), how to build an SEO engine that compounds (the blog), and how to convert free users to paid without a sales team.
These same distribution lessons directly informed how I launched vexp — my current project, a context engine for AI coding agents that got 30 paying users in 15 days using the same playbook.
Built by Nicola Alessi and Lorenzo Pediconilink · Co-founded in 2025
