What it does

This is a Spring Cloud microservices application that proxies calls to OpenAI’s GPT-3.5/4, Baidu’s Ernie Bot, Stable Diffusion, and Midjourney. It exposes chat and text-to-image generation through web, Android, iOS, and H5 frontends. The repo includes a live demo, Bilibili video tutorials, and a WeChat-based login flow.

The interesting bit

The project is built on top of BladeX 3.7.1, an existing Chinese low-code platform — so much of the scaffolding (admin UI, user management, payment-adjacent “points” system) comes pre-assembled. The actual “AI” is entirely third-party API calls; the value is in the integration plumbing and the polished multi-client packaging.

Key highlights

• Supports GPT-3.5, GPT-4, Baidu Wenxin Yiyan, Stable Diffusion, and Midjourney from one backend
• Spring Cloud microservices architecture (Nacos, Gateway) with a Spring Boot monolith variant also available
• Multi-platform clients: web, iOS App Store app, Android APK, WeChat mini-program/official account
• Includes “AI toolbox” presets for education, encyclopedia, life, and entertainment prompts
• Community gallery for sharing generated images with heat-based ranking and poster generation
• Points/credit system via daily check-in for image generation access

Caveats

• README is entirely in Chinese; no English documentation visible
• Heavy reliance on external Chinese CDN assets (oss.mj.ink) that may be slow or blocked outside mainland China
• The “knowledge graph” and “programming assistant” claims in the description are aspirational marketing; the actual implementation appears to be standard prompt-based API calls
• No visible tests, CI configuration, or license file in the provided README excerpt

Verdict

Worth a look if you need a reference architecture for wrapping multiple Chinese and Western AI APIs in a Spring Cloud stack with mobile clients. Skip it if you want a clean, minimal SDK or if you need English-language support and documentation.