Supabase

Overview

Supabase is a managed PostgreSQL platform that does what Firebase does, but without the proprietary walls. You get a real relational database (not a document store), real-time subscriptions, built-in authentication, file storage, and vector embeddings for AI search—all exposed via REST and WebSocket APIs. Who actually uses this: founders building SaaS products where data relationships matter (CRM, project management, marketplace), teams shipping AI features that need semantic search, and anyone who's watched a Firebase bill climb and decided to own their infrastructure instead. The operator scenario: You're three weeks from launch. You need auth that doesn't require a separate service, a database you can query with SQL (not Firebase's limited query language), and the ability to add vector search for your AI feature without ripping out your backend. Supabase handles all three. You sign up, get a Postgres instance, connect your app, and start building. No infrastructure team needed. Real workflow: Create a table in the dashboard or via SQL, enable real-time on specific columns, set row-level security policies (this is the approval gate—who sees what), then subscribe from your frontend. When data changes, clients get notified instantly. This works for collaborative features, live dashboards, and multiplayer experiences. Vector storage is native to Postgres (via pgvector extension). You can store embeddings, search by similarity, and build semantic search or recommendation features without a separate vector database. This matters if you're building AI features and want to avoid managing multiple databases. Pricing is consumption-based: you pay for database size, egress, and compute. Free tier includes a small Postgres instance suitable for prototypes. Paid tiers scale with your usage. No surprise enterprise licensing—you see what you'll pay before you commit. Honest limits: Supabase is still younger than Firebase. The ecosystem is smaller (fewer third-party integrations), and some edge cases require you to understand Postgres (row-level security policies, for example, are powerful but have a learning curve). If you need real-time collaboration on deeply nested documents, Firebase's Firestore might feel more natural. If you need Supabase's power but want zero database knowledge, you'll need a developer. Alternatives: Firebase if you want simplicity and don't mind lock-in. PlanetScale if you prefer MySQL. Railway or Render if you want to manage your own Postgres but want hosting. Neon if you want serverless Postgres without the Supabase layer. The real win: You own your schema. You can write complex queries. You can migrate to self-hosted Postgres tomorrow if Supabase disappears. That's not hype—that's the actual operator advantage.