Supabase Review

Supabase provides a backend around a Postgres database you can take with you. Authentication supports email and password, magic links, OAuth providers, and row-level security. APIs are generated from your schema for REST and GraphQL, realtime subscriptions run over websockets, Edge Functions run TypeScript close to users, storage has access policies tied to auth, and pgvector enables embeddings for AI search. The dashboard includes a table editor, SQL editor, and a visual row-level-security policy builder. OpenAlternative notes the cloud holds SOC 2 Type 2, HIPAA, and ISO 27001 certifications.

Docker with Docker Compose is the recommended self-hosting route, and community projects exist for Kubernetes (an Operator or Helm chart) and for observability, which Supabase does not maintain. The stack runs Studio, an API gateway (Envoy by default, with Kong as an alternative), Auth, PostgREST, Realtime, Storage, imgproxy, postgres-meta, Postgres, Edge Runtime, Supavisor, and optionally Logflare and Vector. Minimum requirements are 4 GB of RAM, 2 CPU cores, and a 40 GB SSD, with 8 GB, 4 cores, and 80 GB recommended. You run a Linux quick-start script or clone the repository, generate secrets with the provided scripts, set SUPABASE_PUBLIC_URL, API_EXTERNAL_URL, and SITE_URL, and start with sh run.sh start, then open Studio on port 8000 behind basic auth. Before starting you must change POSTGRES_PASSWORD, DASHBOARD_PASSWORD, the JWT and API keys, and several encryption keys. You provide your own HTTPS through a reverse proxy (important for OAuth) and your own SMTP server, and storage defaults to local disk or can point at S3-compatible backends. Supabase publishes stable Compose snapshots roughly monthly, and update.sh merges them into your files.

The dashboard is a major draw: create tables, write SQL, and define auth policies visually, then use client libraries from your app. Self-hosted Studio manages a single project with no support for multiple organizations or projects, and most settings are controlled through environment variables. This section reflects the official docs, pricing page, and repository; we did not run the self-hosted stack.

• Auth with email, magic links, OAuth, and row-level security
• Auto-generated REST and GraphQL APIs from your schema
• Realtime subscriptions over websockets
• Edge Functions running TypeScript, plus file storage with access policies
• pgvector for embeddings and AI search
• Dashboard with a table editor, SQL editor, and visual RLS policy builder
• Self-hosting through Docker Compose, with community Kubernetes projects

• Built on standard Postgres, so your data, SQL, and extensions stay portable
• A complete backend in one product with a generous free tier
• Apache-2.0 license and a huge community, with about 111,000 stars and daily commits
• Honest self-hosting docs that list exactly what you take on

• Self-hosted instances run a single project, and branching, advanced metrics beyond logs, managed backups and point-in-time recovery, and the Management API are not available
• You handle HTTPS, SMTP, secrets, high availability, backups, and monitoring, and the self-hosting docs do not cover backups beyond noting the data volumes
• Logs and analytics are off by default and need an override file, which raises resource needs
• On the cloud, free projects are paused after one week of inactivity and capped at 2 active projects, and compute is billed separately from the plan price
• Self-hosting is community-supported unless you are an enterprise customer

Cloud plans: Free ($0) includes a 500 MB database, 50,000 monthly active users, 1 GB of file storage, 5 GB of egress, and 2 active projects, with projects paused after a week of inactivity. Pro starts at $25/month with 100,000 MAUs, 8 GB of disk per project, 100 GB of file storage, and daily backups kept for 7 days. Team starts at $599/month and adds SOC 2 and ISO 27001, dashboard SSO, and 14-day backups. Enterprise is custom with SLAs and premium support. Pro and Team include $10/month in compute credits, which cover one Micro instance.

Firebase is Google's backend platform with databases, authentication, and hosting, built around proprietary NoSQL stores and tied to Google Cloud. Supabase offers a similar developer experience on open-source Postgres, which means relational data, SQL, and the ability to self-host or leave. Firebase has stronger mobile tooling and analytics, while Supabase wins on openness, SQL, and portability. If your data is naturally relational, Supabase is usually the better fit.

• Web and mobile developers who want a full backend without lock-in
• Teams that prefer SQL and relational data
• Projects that need auth, storage, realtime, and vector search in one place
• Companies that want an exit path to self-hosting

• Teams deeply invested in the Google and Firebase mobile ecosystem
• Anyone who wants a self-hosted backend with minimal operations
• Workloads that need a non-relational data model