介紹 CedroDB：以 Rust 開發、供 Python、Node.js 及 AI Agent 使用的開源嵌入式資料庫

🚀 Introducing CedroDB — An Open-Source Embedded Database Built in Rust

I've been developing CedroDB, an experimental, local-first database engine written in Rust, with native bindings for Python, Node.js, and TypeScript.

The goal is to simplify how applications manage persistent data, searchable knowledge, AI agent memory, and recoverable background workflows — all without requiring an external database server.

🔧 Key Features

JSON Document Storage: Store and manage structured data with atomic transactions.
Full-Text Search: BM25-based search for efficient document retrieval.
Vector & Hybrid Search: Combine semantic similarity with text-based retrieval.
AI Agent Memory: Persistent checkpoints with revision checks to support recoverable agent workflows.
Durable Job Queues: Persistent jobs, retries, and recovery mechanisms.
Changefeeds: Track database changes for local integrations and event-driven applications.
Backups & Integrity: Backup and verification capabilities.
Native SDKs: Integrate directly with Python and Node.js/TypeScript.

⚡ Quick Installation

Node.js / TypeScript
npm install cedrodb

Python
pip install cedrodb

💡 Why CedroDB?

Modern AI applications frequently rely on multiple components to manage persistent state, document retrieval, vector search, and background processing.

CedroDB explores bringing these capabilities together within a single embedded database engine.

It is designed for experimentation with:

AI agents and local memory.
Offline-first applications.
Developer tools and automation.
Local knowledge retrieval.
Recoverable background workflows.

🧪 Current Status

CedroDB is currently experimental (v0.9.0) and under active development.

The engine currently operates with a single owning process, keeps active documents and indexes in memory, and does not support SQL, MVCC, or replication.

It is not intended to replace mature databases such as SQLite or PostgreSQL. Instead, it explores a different approach to combining persistence, retrieval, and agent-oriented workflows.

🔗 Explore CedroDB

🌐 Official Website
https://www.ltavarez.me/cedrodb

💻 GitHub — Source Code
https://github.com/uppy19d0/cedrodb

📦 npm — Node.js / TypeScript
https://www.npmjs.com/package/cedrodb

🐍 PyPI — Python
https://pypi.org/project/cedrodb/

💬 Feedback & Contributions

CedroDB is open source, and community feedback is highly appreciated.

I'm particularly interested in feedback on database architecture, API design, performance, reliability, and practical use cases.

What features would you expect from an embedded database designed for AI agents and local-first applications?

⭐ If you find the project interesting, consider starring the repository on GitHub.

Contributions, suggestions, and technical discussions are welcome!