This project offers a novel approach to backend development by leveraging Large Language Models (LLMs) to infer and execute business logic, eliminating the need for traditional coding. It targets developers seeking rapid prototyping and flexible backend iteration, enabling dynamic API endpoints that adapt to user instructions and persistent state.

How It Works

The system operates by taking an initial LLM instruction defining the backend's purpose and a JSON blob representing its state. When an API call is made, the LLM receives the prompt, current state, and the call itself. It then infers the necessary business logic, generates a response, and updates the persistent state. This allows for natural language or pseudo-function call-based interactions, abstracting away explicit backend route definitions and CRUD operations.

Quick Start & Requirements

• Requires an LLM API key (e.g., OpenAI).
• Setup involves providing initial instructions and state.
• Further details and demonstration can be found at backend-gpt.com.

Highlighted Details

• Enables frontend iteration without predefined backend schemas.
• Handles API endpoint variations and misspellings through LLM inference.
• Offers serverless-like performance with minimal cold start due to state-based execution.

Maintenance & Community

• A Discord community is available for support and discussion: https://discord.gg/3ASBTJWgGS.

Licensing & Compatibility

• License information is not explicitly stated in the README.

Limitations & Caveats

The project's effectiveness is heavily dependent on the LLM's inference capabilities and may be limited to simpler applications or specific types of state management. The persistent state is currently limited to 1KB.