butterbase-ai/butterbase

What it solves

Butterbase is an AI-native, open-source backend-as-a-service (BaaS) designed to provide the building blocks for AI-driven applications without vendor lock-in. It eliminates the need for writing repetitive "glue code" by providing a unified infrastructure for data, compute, and AI capabilities.

How it works

Butterbase operates as a runtime data plane that integrates several core services:

• Data Management: Uses a three-plane Postgres architecture (control, runtime, and data planes) to handle platform metadata, runtime tables, and isolated per-app user data with Row-Level Security (RLS).
• Compute: Provides serverless TypeScript functions via the Deno runtime, stateful Durable Objects for long-running agents or multiplayer features, and Edge SSR for frontend hosting.
• AI Integration: Includes an AI gateway for chat and embeddings, a managed RAG (Retrieval-Augmented Generation) system for document ingestion and semantic search, and a Model Context Protocol (MCP) server that exposes backend capabilities as tools for AI agents.
• Infrastructure: Integrates S3/R2-backed file storage, a regional Key-Value store, and WebSocket-based realtime subscriptions.

Who it’s for

Developers building AI-driven applications who want a comprehensive, self-hostable backend that integrates AI tools, database management, and serverless compute in a single package.

Highlights

• MCP Server: Exposes all backend capabilities as tools, allowing AI agents to operate the backend directly.
• AI Gateway & RAG: Built-in support for LLM routing, embeddings, and semantic search.
• Durable Objects: Stateful actors for managing complex agent states or real-time interactions.
• Postgres-backed: Strong data isolation via RLS and declarative schemas.
• Claude Code Plugin: Includes a set of guided skills to accelerate agentic app building.