How nosqlnull/chatgpt-sparkai works
Project type and purpose
• This is a full-stack, modular AIGC (AI-generated content) web platform (“SparkAi系统”) with both B-end (user facing) and C-end (admin) interfaces.
• Purpose: provide one-stop AI services (chat, image/video generation, multimodal analysis, TTS/voice, plugin-based extensions, billing, user/admin management).
File structure
• Top-level:
• README.md – project overview, feature list, architecture.
• SystemPreview/ – static screenshots; not code.
• Implied code directories (per README):
• frontend/ – Vite + Vue3 + TypeScript + TailwindCSS for public site.
• admin/ – Vite + Vue3 + Element-Plus for management console.
• backend/ – NestJS (Node.js) monolithic server with multiple modules.
• config/, scripts/, docker/ or deploy files (for Baota/Docker).
README for additional insights
• Describes tech stack: Vite, Vue3, NaiveUI, Tailwind; NestJS; MySQL 5.7+; Redis; local or cloud storage (OSS/COS/Chevereto).
• Lists dozens of functional modules (auth, AI model adapters, billing, plugin system, TTS, multimodal, video, admin dashboards, analytics, distribution, card system, menus, notifications, etc.)
Final advice
• Network boundaries (e.g., DMZ, internal network).
• Asynchronous flows (e.g., video generation via job queue).
• Keep it detailed but avoid overcrowding—group related modules under a “Backend” umbrella and expand only key modules.