wrtnlabs/autobe

Describe your backend requirements in natural language through AutoBE's chat interface.

AutoBE will analyze your requirements and build the backend application for you. The generated backend application is designed to be 100% buildable by AI-friendly compilers and ensures stability through powerful e2e test functions.

With such AutoBE, build your first backend application quickly, then maintain and extend it with AI code assistants like Claude Code for enhanced productivity and stability.

AutoBE will generate complete specifications, detailed database and API documentation, comprehensive test coverage for stability, and clean implementation logic that serves as a learning foundation for juniors while significantly improving senior developer productivity.

To use AutoBE, clone the repository and run the playground application locally. This allows you to chat with AutoBE's AI agents, manage multiple sessions, and use various LLM providers including local models like qwen3.5-397b-a17b.

After installation, the playground will be available at http://localhost:5713. You can interact with AutoBE through a chat interface - simply describe what you want to build, and AutoBE will generate the backend application for you.

Here's an example conversation script that guides AutoBE to create an "Economic/Political Discussion Board":

The playground includes a replay feature at http://localhost:5713/replay/index.html where you can view chat sessions from the AutoBE development team's testing and benchmarks.