Software & Guides/Guide/Explainer

What Is Supabase? The Open-Source Firebase Alternative Explained

Supabase bundles a Postgres database, auth, storage, and serverless functions into one open-source backend platform. Here is how it works and what it costs.

The Supabase logo and wordmark on a dark green background
Image: Supabase.

Supabase is an open source Postgres development platform: a hosted (or self-hosted) backend that bundles a full Postgres database with authentication, file storage, auto-generated APIs, realtime sync, serverless functions, and AI/vector search into one project. Its own tagline sums up the pitch: "Build in a weekend. Scale to millions." You get a dedicated Postgres database for every project, plus the pieces most apps need on top of it, without stitching together five separate services yourself.

Supabase in short: an open source alternative to Firebase, built on standard Postgres instead of a proprietary database. Core products are Database, Auth, Storage, Edge Functions, Realtime, and Vector/AI search. A Free tier exists ($0/month); paid plans start at $25/month (Pro). The core stack is Apache-2.0 licensed and can be self-hosted.

What is Supabase?

Supabase describes itself as "the Postgres development platform," and its own GitHub repository says more plainly that the project is "building the features of Firebase using enterprise-grade open source tools." That comparison is the fastest way to understand it: where Google's Firebase gives you a managed NoSQL database (Firestore) plus auth, storage, and functions behind a proprietary system, Supabase gives you the same category of features but built on top of standard, open source PostgreSQL — a database you can query with regular SQL, inspect with any Postgres client, and move off the platform without a rewrite.

Every Supabase project is, underneath, a real Postgres database. Supabase layers tooling on top of it: a Table Editor and SQL Editor in the dashboard, auto-generated APIs, and a set of additional products (auth, storage, functions, realtime, vector search) that all read and write against that same database rather than separate, disconnected systems.

How Supabase works

When you create a new Supabase project, you get a dedicated Postgres instance immediately. From there, Supabase automatically generates APIs from your database schema: a REST API (powered by the open source PostgREST), a GraphQL API (powered by pg_graphql), and realtime subscriptions over WebSockets. Define a table in SQL, and an API for it exists moments later — no separate backend code required for basic CRUD operations.

Authorization is handled through Postgres' own Row Level Security (RLS), rather than a bolt-on permissions layer. That means access rules live as SQL policies attached to your tables, so the same rule applies no matter which API (REST, GraphQL, or a client library) is used to reach the data.

The core building blocks

Supabase is organized as a set of products that share one Postgres database. According to Supabase's own site and documentation, the main pieces are:

  • Database — a dedicated, portable Postgres database with auto-generated REST and GraphQL APIs, a Table Editor, a SQL Editor, and more than 40 preinstalled extensions.
  • Auth — built-in user management supporting more than 20 social sign-in providers, email/password, phone/OTP, and magic links, with authorization enforced through Postgres Row Level Security.
  • Storage — S3-compatible object storage with a global CDN and image transformations, split across file, analytics (Apache Iceberg), and vector bucket types.
  • Edge Functions — globally distributed TypeScript/Deno serverless functions with Node.js compatibility and access to the npm ecosystem, for logic that needs to run outside the database.
  • Realtime — WebSocket-based sync that covers three patterns: listening for database changes, tracking presence (who's online), and broadcasting arbitrary messages between connected clients.
  • Vector / AI — AI tooling built on the pgvector Postgres extension, for storing, indexing, and querying embeddings alongside the rest of an app's transactional data, which is what makes Supabase a common backend choice for retrieval-augmented generation (RAG) apps. Pandromeda's own vector database explainer and RAG explainer cover that pattern in more depth.

You don't have to use every product. Supabase's pitch is that unused parts of the platform cost nothing — a team that only wants the Postgres database and auto-generated APIs can skip Auth, Storage, or Edge Functions entirely.

Open source and self-hosting

Supabase's core stack is open source under the Apache-2.0 license. Its GitHub repository shows the project is built from several existing open source tools wired together: Postgres itself, an Elixir-based Realtime server, PostgREST for the REST API, GoTrue for authentication, a Storage API, pg_graphql for GraphQL, postgres-meta, and Envoy as a proxy. Because every one of those pieces is independently open source, a team can self-host the entire stack — on their own servers or any cloud — rather than depend on Supabase's hosted platform. Official client libraries exist for JavaScript/TypeScript, Flutter, Swift, and Python, with community-maintained libraries covering other languages.

That self-hostability is the biggest practical difference from Firebase, which has no equivalent open source, run-it-yourself version. A Supabase project built on the hosted platform can, in principle, be exported and run independently since it's just Postgres plus open components underneath — reducing the vendor lock-in that comes with proprietary backend platforms.

What's running underneath Supabase

Supabase's own architecture documentation lays out exactly which open source tools sit behind each product, and the list explains why the platform is self-hostable rather than a black box. Postgres is the core — Supabase says it deliberately doesn't abstract the database away, so a project's data is always queryable with ordinary SQL and standard Postgres tooling. PostgREST turns that database into a RESTful API automatically; paired with the pg_graphql extension, the same schema also gets a GraphQL API with no extra code. GoTrue, the auth service, issues JWT access tokens and integrates directly with Postgres Row Level Security, so authentication and authorization share one source of truth. The Storage API is S3-compatible object storage that keeps file metadata inside the same Postgres database rather than a separate system. A tool called postgres-meta exposes a RESTful API for managing the database itself — fetching table definitions, adding roles, running queries — which is what powers the dashboard's Table Editor. Supavisor is a multi-tenant connection pooler that lets a single Postgres instance handle many simultaneous client connections efficiently, and Envoy acts as the API gateway that routes incoming requests to the right service. Studio is the web dashboard that ties all of this together into one interface. Each of these components can also be used independently outside Supabase, since every one of them is its own open source project with its own API.

Supabase pricing

Supabase publishes four tiers on its pricing page. Figures below are what the page lists as of this writing; check the live page before budgeting, since usage-based add-ons can change the final bill.

PlanPriceKey limits
Free$0/month50,000 monthly active users, 500 MB database, 1 GB file storage, 5 GB egress, max 2 active projects, pauses after 7 days of inactivity
ProFrom $25/month100,000 MAUs, 8 GB disk, 250 GB egress, 100 GB storage, daily backups (7 days), $10/month compute credit, projects don't pause
TeamFrom $599/monthPro's quotas plus SOC 2 and ISO 27001 compliance, SSO for the dashboard, 14-day backups, 28-day log retention, HIPAA available as a paid add-on
EnterpriseCustomTechnical Account Management, uptime SLAs, AWS PrivateLink support, 24/7 premium support, a private Slack channel

Free-tier projects pause automatically after a week of low activity; paid plans remove that pause. For a team comparing Supabase's pricing directly against Google's Firebase, Pandromeda's Supabase vs. Firebase comparison lays the two models side by side.

Getting started with Supabase

To start a project, sign up at supabase.com and create a new project, which provisions a dedicated Postgres database in one of Supabase's supported regions. From there:

  • Install the CLI with npm install supabase --save-dev, then authenticate with npx supabase login (this opens a browser window).
  • Define tables through the dashboard's Table Editor or directly in SQL via the SQL Editor — both generate REST and GraphQL endpoints automatically.
  • Connect a client library (JavaScript/TypeScript, Flutter, Swift, or Python) to call those auto-generated APIs from an app.
  • Supabase also publishes agent skills (npx skills add supabase/agent-skills) and an MCP server at mcp.supabase.com for AI coding tools that support the Model Context Protocol.

Supabase and AI coding tools

Supabase has leaned into AI-assisted development directly on its homepage: it publishes a set of installable "agent skills" (npx skills add supabase/agent-skills) meant to teach AI coding assistants how to work with a Supabase project correctly, and it runs its own Model Context Protocol (MCP) server at mcp.supabase.com, authenticated through OAuth. MCP is an open standard that lets AI coding tools call out to external services — in this case, letting an AI assistant inspect or modify a Supabase project's schema, run queries, or manage auth settings on a developer's behalf during a coding session, instead of the developer switching to the dashboard by hand. Combined with the built-in pgvector support described above, that positions Supabase as a backend aimed specifically at teams building AI features, not just a general-purpose Firebase substitute.

Who uses Supabase, and why Postgres matters

Supabase is aimed at the same audience Firebase has long served — indie developers and startups who want a working backend without standing up their own infrastructure — plus teams building AI applications that need vector search alongside regular transactional data. Building on Postgres rather than a proprietary database is the main reason developers cite for choosing it: Postgres is a decades-old, heavily documented relational database with a huge extension ecosystem (40+ extensions ship with Supabase projects), which means SQL skills, existing ORMs, and existing Postgres tooling carry over directly instead of requiring a new, platform-specific query language.

Supabase is also SOC 2 Type 2 compliant and runs in more than 16 regions, which matters for teams with data-residency or compliance requirements that a side-project backend wouldn't normally need to satisfy.

Common points of confusion

A few things trip people up when they first look at Supabase. It is not a Postgres hosting company in the narrow sense — the database is the foundation, but the product is the whole bundle of Auth, Storage, Realtime, Edge Functions, and Vector search layered on top. It is also not literally built from Firebase's code; "the Firebase alternative" is a positioning comparison, not a technical lineage. And open source does not mean the hosted platform is free to run at scale — the Free tier is explicitly capped and intended for small or early-stage projects, which is why Supabase's own TLS and infrastructure changes (like the cipher update covered in Pandromeda's recent Supabase TLS report) matter even to free-tier users running production traffic.

What to do next

If the goal is a quick prototype, the Free tier is enough to build and test a full app — database, auth, storage, and APIs included — before paying anything. Teams that need the project to stay always-on, or that expect to exceed the Free tier's 50,000 MAU or 500 MB database limit, should budget for the Pro plan starting at $25/month. Teams with compliance requirements (SOC 2, ISO 27001, HIPAA) should look at the Team tier, and organizations needing guaranteed uptime or private networking should talk to Supabase's sales team about Enterprise pricing. Anyone evaluating Supabase against a managed NoSQL alternative should start from the fact that it is Postgres at its core — the deciding question is usually whether a relational, SQL-first database fits the app better than a document database, not which platform has more features.

Frequently asked questions

Is Supabase free to use?

Yes. Supabase's Free tier costs $0/month and includes up to 50,000 monthly active users, a 500 MB database, and 1 GB of file storage, though free projects pause after a week of low activity.

Is Supabase open source?

Yes. Its core stack, including Postgres, PostgREST, GoTrue, Realtime, and Storage, is released under the Apache-2.0 license and can be self-hosted.

Is Supabase built on Firebase?

No. Supabase is an independent, Postgres-based platform that positions itself as an open-source alternative to Firebase, not code derived from it.

What database does Supabase use?

Every Supabase project is a dedicated PostgreSQL database, with REST and GraphQL APIs generated automatically from its schema.

How much does Supabase Pro cost?

Supabase's Pro plan starts at $25/month and raises the Free tier's limits to 100,000 monthly active users, 8 GB of database storage and 250 GB of egress.

Can I self-host Supabase?

Yes. Because every component of Supabase's stack is open source, the whole platform can be run on a team's own servers or cloud account instead of Supabase's hosted service.

Sources

More on Supabase →SupabasePostgresBackend as a ServiceOpen SourceFirebase Alternative
Felix Moreau
Written byFelix Moreau

Felix Moreau writes Pandromeda's software coverage and how-to guides. He covers Windows, macOS and Linux updates, the apps people rely on, emulators and developer tools, and turns official documentation into clear, numbered steps that work on the current version.

More from Software & Guides

See all