This project provides a modular, extensible web client for AI chat and image generation services, inspired by a "beehive" architecture. It targets users who want a customizable interface for interacting with various AI models, offering features like dynamic room creation based on "blueprints" and support for services like OpenAI GPT and Midjourney.

How It Works

The system uses a modular design, allowing new AI services and room types to be added by defining new "blueprints." This approach enables significant extensibility, mirroring the honeycomb structure for scalability. The frontend is built with Vite, supporting features like dynamic configuration, authentication, and real-time data streaming for chat responses.

Quick Start & Requirements

• Install dependencies: pnpm install (recommended) or npm install.
• Run locally: pnpm dev.
• Build for production: pnpm build.
• Requires Node.js v16.15.1 and npm v8.11.0 or pnpm v7.5.0.
• Project configuration: Modify .env file for API base URLs.
• Preview site: https://front.aibeehive.icu

Highlighted Details

• Supports OpenAI GPT 3.5 & 4, OpenAI Image, and Midjourney (Imagine, Upscale, Variation, Describe).
• Implements streaming responses for a typewriter-like chat experience.
• Includes dynamic room creation and configuration based on blueprint parameters.
• Planned features include support for NewBing, Wenxin Yiyan, Tongyi Qianwen, and ChatGLM.

Maintenance & Community

The project is actively developed with a roadmap outlining planned features. Users are encouraged to report issues via GitHub issues. Paid support is available for development or deployment assistance.

Licensing & Compatibility

Licensed under the Apache-2.0 license. This license is permissive and generally compatible with commercial use and closed-source linking.

Limitations & Caveats

The project is intended for learning and research and is not recommended for commercial use. The developers are not responsible for any losses incurred from its use. While issues are prioritized, the development team may have limited capacity to answer all user questions due to other commitments.