Schema, database, document, user and dev-server tools for a Rebase backend project.
Copy the AI prompt to install this server into Claude Code, Cursor, or another agent β or use 1-click editor setup below.
One-click editor setup isnβt available for this listing yet β we donβt have a confirmed install command, and weβd rather show nothing than point your editor at the wrong package or host. Follow the projectβs own setup instructions, linked above.
Point it at a database and get a working backend in minutes.
REST, auth, storage, realtime and backups over your own Postgres β then add a
schema-driven admin panel, or don't.
Own your data, own your code. The absolute easiest way to build on PostgreSQL.
Public beta. The API you write against can change in a minor,
with a changelog entry. Your data cannot break quietly.
What that promise rests on, and
which subsystems are ready
Live Demo β’ Documentation β’ Features β’ Discord
Rebase is a developer-first, open-source backend-as-a-service built on PostgreSQL β authentication, S3-compatible storage, realtime, backups and auto-generated REST APIs over a database you own. It is self-hosted and agent-native, with a built-in MCP server for AI-assisted development.
The admin panel is the second layer of the same definition: add collection definitions and you get a schema-driven back-office built from them, extensible with your own React components, serverless functions, and scripts.
Three modes, same packages, wired differently β see MODULAR-ARCHITECTURE.md.
| Mode | You get | Comparable to |
|---|---|---|
| BaaS | REST + auth + storage + realtime + backups over your database. No config files, no UI, no React in the dependency tree. | Supabase |
| CMS | BaaS + a schema-driven admin UI built from your collection definitions. | Payload, Directus |
| Full | CMS + Studio (SQL editor, schema visualizer, RLS editor, logs, API explorer). | Supabase + Payload |
pnpm dlx @rebasepro/cli init scaffolds a complete project in seconds.Scaffold a complete, self-hosted Rebase application connected to your database:
That is the whole first run β no database to install, no schema step. With no
DATABASE_URL set, rebase dev starts a managed PostgreSQL (PGlite) that lives
in the project directory, generates the schema from your collections, and creates
the tables at boot.
Read the URLs from its output. rebase dev picks a free port per project
rather than fixed ones, so they differ between projects. PORT and
VITE_API_URL in .env apply to rebase start, not here.
Useful flags: --yes (required when there is no terminal to prompt, such as CI),
--headless (see below), --template <name>, and --install / --no-install
on init; --docker on dev to use the compose Postgres in this project
instead, and --no-db to start nothing β set DATABASE_URL yourself.
To use your own Postgres instead: uncomment DATABASE_URL in .env and run
pnpm run dev again. Nothing else changes β a DATABASE_URL that is set is
never touched, and one pointing anywhere other than this machine is left alone
entirely.
Rebase is modular β take only the parts you want:
That scaffolds a headless backend: REST, auth, storage, realtime and backups over your database, with no collection files and no UI. Tables are served automatically, derived from your schema at boot β change the schema with a migration and the API follows. No React anywhere in the dependency tree.
A fresh headless project has no tables yet, so /api/data/* answers
404 NO_COLLECTIONS until you create some. Point it at a database that already
has tables, or add them with a migration, and they appear.
The three adoption modes β BaaS (like Supabase), CMS (like Payload/Directus), and both together β are described in MODULAR-ARCHITECTURE.md.
An incredibly fast, windowed spreadsheet view to manage your database with inline editing, real-time updates, filtering, sorting, and text search. Switch between multiple view modes:
Define your data models using pure TypeScript collections. Rebase automatically generates your Drizzle ORM schema, handles PostgreSQL migrations, and keeps your live database perfectly in sync using built-in tooling like rebase doctor.
First-class database backups for self-hosted Postgres, built on pg_dump / pg_restore:
rebase db backup --out <path|s3://β¦> β compressed, custom-format dumpsrebase db restore <backup> β confirmation-gated, with --create-db / --target-db to restore into a fresh databaseSee docs/backups.md for the full guide, including security and PITR notes.
Built-in authentication with multiple providers:
Granular role-based access control (RBAC) with customizable permissions per collection, field, and action.
Native S3-compatible file storage with:
A full developer environment built into the admin panel β SQL and JS editors, RLS policies, a schema visualizer, database branches, backups, request logs, cron jobs, a storage browser, an API explorer and API keys. What each tool does β
A full WebSocket engine built into every Rebase backend:
Opt in with offline: true and the client SDK keeps a local database of rows instead of a cache of responses:
No reviews yet β be the first to share how this listing worked for you.
Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.
[](https://allmcps.com/mcp/rebase)<a href="https://allmcps.com/mcp/rebase"><img src="https://allmcps.com/api/badge/rebase?style=directory" alt="Rebase on AllMCPs" /></a>