I built LayerNEXUS, a self-hosted tool that turns messy, denormalized CSV files into clean, normalized SQL schemas and ER diagrams - 100% offline and privacy-first.
Key Features: - Drag & drop CSVs/JSONs (multi-file supported) - Auto-detects entities, relationships, and infers 3NF structure - Outputs SQL for PostgreSQL, MySQL, SQLite, Oracle, and SQL Server - Instant ERD preview - AI-powered schema refinement (bring your own OpenAI key)
Why I built it: I was tired of reverse-engineering messy Excel/CSV datasets and manually drawing ER diagrams. I needed a fast way to convert real-world data dumps into clean, production-grade schemas.
Would love your feedback - bugs, feature ideas, UX thoughts, anything.