butterbase-ai

butterbase: Postgres Backend for AI Applications

Butterbase is an AI-native backend-as-a-service built on Postgres, with auth, storage, functions, realtime, an AI gateway, RAG, and MCP tools. Self-host it with Docker Compose or use the managed cloud for AI-driven apps without proprietary data storage.

NewTypeScriptApache-2.0DatabaseSelf HostedMCP
View on GitHub Compare
Stars
3k
Forks
166
Last commit
7 days ago
Contributors
6
Releases, 12 months
0

Our take

Newless than six months old and changing fast

Health score

45/100

How it is scored
Activity
53
Community
19
Issues
21
Pull requests
85
  • Merges more pull requests than 98% of the projects we track

Butterbase combines a broad Postgres backend with AI and MCP capabilities, but its young, fast-changing codebase and highly concentrated contributions make it a higher-adoption-risk choice today.

Good fit if

  • You want Postgres, auth, functions, storage, realtime, and AI features in one backend, with the option to self-host under Apache-2.0.
  • Your application or agent can benefit from MCP access to backend operations, RAG, or declarative agent workflows.
  • You can work with a rapidly changing project and have the capacity to evaluate or adapt its self-hosted runtime.

Look elsewhere if

  • You need managed multi-region orchestration, billing, or upstream AI router integrations in the open-source distribution.
  • Your team cannot accommodate the stated Node.js and Python runtime requirements or the Docker Compose self-hosting workflow.
  • You need a project with contributions spread across maintainers: most code comes from one person, despite recent high commit and merge activity.
All health signals
Last commit2026-11-08 (7 days ago)
Commits, last 90 days100+
Releases, last 12 months0
Contributors6 (top contributor: 99% of commits)
Issues closed, last 90 days0+
Pull requests merged, last 90 days97 (typically merged in 0 days)
Project age6 months

Checked on 2026-11-15 with the GitHub API.

Overview

Butterbase is an AI-native, open-source backend-as-a-service built on Postgres. It gives each app its own database and combines data, auth, storage, serverless functions, realtime, AI features, and an MCP server so applications and agents can use backend capabilities without stitching together separate services. The project is available as a self-hosted Docker Compose stack or through Butterbase's managed cloud.

It is aimed at developers building AI-driven applications who want Postgres, row-level security, and agent-facing tools in one backend. The repository contains the runtime data plane; managed-cloud orchestration, billing, and some AI provider integrations are separate and not included.

Key Features

  • Per-app Postgres databases with schema management, migrations, generated REST APIs, and row-level security.
  • Auth, service keys, file storage, key-value storage, and audit logs.
  • TypeScript serverless functions running on Deno, plus realtime subscriptions and durable per-key actors.
  • An AI gateway, document ingestion and search for RAG, and declarative multi-step agent workflows.
  • MCP tools over HTTP or stdio, allowing agents to operate backend capabilities.
  • Frontend hosting and edge SSR deployment, including support for Next.js, Remix, and Astro handlers.
  • TypeScript SDK and CLI for application and backend workflows.

Use Cases

  • A startup building an AI product that needs a Postgres-backed API, user auth, and file storage without assembling each backend component separately.
  • A team giving an AI agent controlled access to application data and operations through MCP tools or function-based workflows.
  • Developers prototyping a live application with realtime updates, serverless functions, and per-user data access enforced by RLS.
  • Teams that want to self-host their backend stack, or evaluate the same runtime data plane used by Butterbase's managed offering.

What you need

Detected in the repository

  • Node.js >=22.0.0 (from package.json)
  • Python >=3.14 (from pyproject.toml)
  • A test suite

License in plain words

Apache-2.0permissive

  • Commercial use: yes
  • Modify and redistribute: yes
  • You must keep: the license, the NOTICE file and a note of your changes
  • Share your changes: no
  • Includes an explicit patent grant from the contributors:

A summary, not legal advice: the LICENSE file is what applies.

Getting Started

The self-host guide requires Docker, Node.js 22+, and npm. Clone the repository with its submodules, then start the local stack:

git clone --recurse-submodules https://github.com/butterbase-ai/butterbase.git
cd butterbase
npm ci
cp .env.example .env
docker compose -f docker-compose.local.yml up -d

Database migrations and a development-user seed are separate steps. See the README and SETUP.md for the complete setup.

Considerations

  • The project is new and changing quickly. There have been no releases in the past year, and the README identifies v0.3.0 as the latest release.
  • Contributor activity is concentrated: most code comes from one person. Recent commit and pull request activity is high, but the measured data reports no issues closed in the last 90 days.
  • Self-hosting involves a multi-service stack and database setup. The project requires Node.js 22 or later and Python 3.14 or later; Docker Compose is used in the documented setup.
  • The open-source distribution does not include managed multi-region orchestration, billing and quota logic, or upstream AI router adapters. Self-hosters may need to supply implementations through the documented interfaces.
  • The repository is licensed under Apache-2.0. Review the license and its obligations when redistributing or modifying the software.
  • The README describes the managed data plane as production-tested, but also calls the open-source distribution young and invites self-hosting feedback, so evaluate the OSS deployment for your own operational needs.

Found this useful?

Share it with someone who would like butterbase.