AI-Powered Gaming Memory Companion SaaS with Steam Integration

Ludora is an AI-powered memory companion for gamers who juggle too many games to remember where they left off. Come back to a game after days, weeks, or months, ask "where was I?", and get a real answer, grounded in your own session history, notes, and screenshots via semantic search, not a generic guess.

It's built as three connected apps sharing one Supabase backend: a Next.js web app, a Chrome extension, and a Windows desktop companion. Screenshot capture and AI vision analysis let you drop in an in-game screenshot and get it understood and answered against; automatic session summarization and memory extraction mean you don't have to write anything down yourself. A Steam integration imports real playtime and achievement data, and a zero-cost "what should I play tonight?" picker uses a deterministic heuristic instead of another metered AI call.

The problem it solves: people who rotate through 5-10 games lose track of quest state, builds, and where a save file left off. Wikis and personal notes don't help. They're generic and require manual upkeep. Ludora is grounded in your own actual play history and updates itself automatically.

It's a complete, working product, live in production, not a prototype with real AI integration (Gemini + Voyage AI, Anthropic supported as a drop-in swap), real Paddle billing (subscriptions + a one-time Lifetime tier), a full Supabase schema with RLS, and a genuine test suite. Built solo, part-time, and fully functional today.