Multimodal AI client (React + Vite + Tailwind, Express + MySQL) with chat, vision, and file analysis.
多模态 AI 客户端(React + Vite + Tailwind,Express + MySQL),支持聊天、视觉与文件分析。
📚 Full guides / 完整说明:
- 🇨🇳 README.zh-CN.md
- 🇺🇸 README.en.md
npm install
npm run server # backend http://localhost:3001
npm run dev # frontend http://localhost:5173- 🧠 Multimodal: chat, vision, file/folder analysis with unified Gemini model selector
多模态:聊天、视觉、文件/文件夹分析,统一模型下拉选择 - 📂 Sessions: create/rename/delete, pin to top, draggable sidebar, manual titles preserved
会话:创建/重命名/删除,置顶,侧栏可拖拽,手动标题不被覆盖 - ⚡ Message actions: continue/retry/summarize/translate; Enter send, Shift+Enter newline
消息操作:继续/重试/总结/翻译;Enter 发送,Shift+Enter 换行 - 📝 Prompt templates: built-in & custom, persisted locally, quick insert
提示词模板:内置+自定义,持久化存储,快捷插入 - 📥 Export: one-click download conversation as Markdown (not clipboard)
导出:一键下载会话 Markdown 文件(本地下载) - 🌐 Status: DB & network health/latency, region/IP/ISP card, collapsible panel
状态:数据库/网络健康与延迟,地区/IP/运营商卡片,可折叠面板 - 💰 Insight: latency, token counts, estimated cost per reply
成本视图:展示延迟、Token 数量与估算成本
- Frontend: React 18, Vite, Tailwind, Lucide
前端:React 18、Vite、Tailwind、Lucide - Backend: Node.js, Express, MySQL (auto-init tables)
后端:Node.js、Express、MySQL(自动建表) - AI: Google Gemini SDK (config via
VITE_API_KEY)
AI:Google Gemini SDK(通过VITE_API_KEY配置)