MiniChat — Enterprise AI Chatbot

A production-ready, cross-device AI chatbot powered by MiniMax M2.7 — text, vision, speech, and search in one unified subscription.

Features

Capability	Description
M2.7 Chat	Streaming token-by-token responses with full conversation context
Web Search	Real-time web search via MiniMax MCP `web_search` tool
Image Understanding	Analyze images with `understand_image` MCP tool
Text-to-Speech	40+ voices, 7 emotions, voice cloning via `speech-2.8-hd`
Image Generation	Photorealistic image generation via `image-01` model
Cross-Device Sync	Conversations sync in real-time across all your devices
Offline PWA	Works offline with queued messages and background sync
Dark Theme	Enterprise-grade dark UI on all platforms

Screenshots

Note

Add screenshots by placing images in docs/screenshots/ and referencing them here.

┌─────────────────────────────────────────────────────────────┐
│  MiniChat — M2.7 Enterprise Assistant                        │
│                                                             │
│  [Chat List]        [Conversation]         [Settings]        │
│                                                             │
│  ○ New Chat         ┌─────────────────────┐                  │
│                     │ Assistant          │                  │
│  ▌ Project Setup    │ What's the best    │                  │
│    "How do I set..." │ architecture for.. │                  │
│                     │                     │                  │
│  ▌ Code Review      │ ┌─────────────────┐│                  │
│    "Can you check..." │ I'd recommend a ││                  │
│                     │ microservices...  ││                  │
│  ▌ Debug Help       │                     │                  │
│    "Why is my API.." │ 🔍 Searching...   ││                  │
│                     └─────────────────────┘                  │
│                     │ Type a message...  [📎] [▶] │        │
└─────────────────────────────────────────────────────────────┘

Architecture

┌─────────────────────────────────────────────────────────────┐
│                        Clients                                │
│  ┌────────┐  ┌────────┐  ┌────────┐  ┌────────┐           │
│  │  iOS   │  │Android │  │   Web  │  │Desktop │           │
│  │  React │  │ React  │  │  Vite  │  │  Tauri │           │
│  │  Native│  │ Native │  │   PWA  │  │  + Rust│           │
│  └────┬───┘  └────┬───┘  └────┬───┘  └────┬───┘           │
└───────┼──────────┼──────────┼──────────┼────────────────────┘
        │          │          │          │
        └──────────┴──────────┴──────────┘
                      │ HTTPS / WSS
                      ▼
┌─────────────────────────────────────────────────────────────┐
│                    API Gateway (Fastify)                     │
│  JWT Auth │ Rate Limiter │ WebSocket │ SSE Streaming       │
└────────────────────────────┬────────────────────────────────┘
                             │
        ┌────────────────────┼────────────────────┐
        ▼                    ▼                    ▼
┌───────────────┐  ┌────────────────┐  ┌─────────────────────┐
│ MiniMax M2.7  │  │  MCP Tools     │  │  Multimodal APIs    │
│  (Streaming)  │  │ web_search     │  │  TTS │ Image Gen   │
│               │  │ understand_img │  │  Music │ Video       │
└───────────────┘  └────────────────┘  └─────────────────────┘

Tech Stack

Frontend

Platform	Framework	Key Libraries
iOS / Android	React Native + Expo	Zustand, TanStack Query, expo-* modules
Web	Vite + React + PWA	Tailwind CSS, Workbox, react-markdown
Desktop	Tauri 2.x + Rust	Native window, notifications, OS APIs

Backend

Layer	Technology
Server	Node.js + Fastify 5
Language	TypeScript (strict)
Auth	JWT with httpOnly cookies
Rate Limiting	Sliding window (4500 req / 5 hrs)
Context Store	In-memory (swap for Redis in prod)

MiniMax APIs

API	Endpoint	Used For
M2.7 (Anthropic-compatible)	`https://api.minimax.io/anthropic/v1/messages`	Chat streaming, function calling
Web Search MCP	Tool in M2.7 request	Real-time web queries
Image Understanding MCP	Tool in M2.7 request	Image analysis
TTS	`https://api.minimax.io/v1/t2a_v2`	Voice synthesis
Image Gen	`https://api.minimax.io/v1/image_generation`	Text-to-image

Getting Started

Prerequisites

Node.js 20+
pnpm (or npm)
Rust 1.70+ (for desktop app)
A MiniMax Token Plan subscription (Plus recommended)

1. Clone & Install

git clone https://github.com/YOUR_USERNAME/chatbot-enterprise.git
cd chatbot-enterprise

# Install root dependencies (if any)
# Install each subproject
(cd backend && npm install)
(cd web && npm install)
(cd mobile && npm install)
(cd desktop && npm install)

2. Configure Environment

# backend/.env
MINIMAX_API_KEY=sk-cp-your-key-here
MINIMAX_API_HOST=https://api.minimax.io
JWT_SECRET=generate-a-strong-random-secret-here
PORT=3000

Important

Never commit .env files. Each subproject has a .env.example — copy it and fill in real values.

cp backend/.env.example backend/.env
# Edit backend/.env with your MiniMax API key

3. Start the Backend

cd backend
npm install
npm run dev

# Server running at http://localhost:3000
# Health check: curl http://localhost:3000/health

4. Start the Web App

cd web
npm install
npm run dev

# Opens at http://localhost:5173
# (proxies /api/* and /ws to backend)

5. Run the Mobile App

cd mobile
npx expo start

# Opens Expo Dev Tools
# Press i for iOS Simulator, a for Android Emulator
# Scan QR code with your phone for device testing

6. Build the Desktop App

cd desktop
npm install
npm run tauri dev       # Development
npm run tauri build     # Production build (.exe/.dmg/.AppImage)

Project Structure

chatbot-enterprise/
├── backend/              # Fastify API server
│   ├── src/
│   │   ├── index.ts           # Entry point, CORS, JWT, WebSocket
│   │   ├── routes/
│   │   │   ├── auth.ts         # Register, login, JWT
│   │   │   ├── chat.ts         # Core streaming chat endpoint
│   │   │   └── conversations.ts # CRUD for conversations
│   │   ├── lib/
│   │   │   ├── minimax.ts      # MiniMax API client (Anthropic-compatible)
│   │   │   ├── rate-limiter.ts  # 4500 req/5hr sliding window
│   │   │   └── conversation-store.ts
│   │   └── types.ts
│   └── .env.example
│
├── web/                 # PWA (Vite + React)
│   ├── src/
│   │   ├── pages/              # Login, Register, Chat, Settings
│   │   ├── components/          # ChatList, MessageBubble, ChatInput...
│   │   ├── stores/             # Zustand (auth + chat state)
│   │   ├── hooks/              # useStreamingChat, useWebSocket
│   │   ├── lib/                # api.ts, websocket.ts
│   │   └── service-worker.ts    # Workbox PWA
│   ├── public/                  # manifest.json, icons
│   └── dist/                   # Built output
│
├── mobile/              # React Native + Expo
│   ├── app/                   # Expo Router pages
│   │   ├── (auth)/             # Login, register screens
│   │   ├── (tabs)/             # Chat list, settings tabs
│   │   └── chat/[id].tsx       # Individual chat
│   ├── components/chat/        # MessageBubble, StreamingText, ToolCallCard
│   ├── stores/                 # Zustand + AsyncStorage persistence
│   ├── hooks/                  # useStreamingChat, useNativeCapabilities
│   ├── lib/                    # api.ts, websocket.ts
│   └── constants/              # Theme, voice catalog
│
└── desktop/             # Tauri 2.x + Rust
    ├── src-tauri/
    │   ├── src/main.rs          # Rust entry + plugins
    │   ├── tauri.conf.json      # Window, plugins, build config
    │   └── Cargo.toml            # Rust dependencies
    └── src/                     # Shared web frontend hooks

API Reference

Authentication

Method	Endpoint	Body	Response
POST	`/api/auth/register`	`{ email, password }`	JWT in httpOnly cookie
POST	`/api/auth/login`	`{ email, password }`	JWT in httpOnly cookie
POST	`/api/auth/refresh`	—	New JWT
POST	`/api/auth/logout`	—	Clears cookie

Conversations

Method	Endpoint	Body	Response
GET	`/api/conversations`	—	`Conversation[]`
POST	`/api/conversations`	`{ title? }`	`Conversation`
GET	`/api/conversations/:id`	—	`Conversation` with messages
DELETE	`/api/conversations/:id`	—	`204 No Content`

Chat

Method	Endpoint	Body	Response
POST	`/api/chat/message`	`{ conversationId, content, attachments? }`	`text/event-stream`

Streaming events:

event: token
data: {"content": "Hello"}

event: tool_call
data: {"name": "web_search", "input": {"query": "..."}}

event: done
data: {"usage": {"input_tokens": 100, "output_tokens": 200}}

Rate Limiting

Your Plus plan quota: 4,500 requests / 5 hours (sliding window, per user).

The backend enforces this server-side. When exceeded:

{ "error": "Rate limit exceeded", "retryAfter": 1234 }

MiniMax Token Plan — What You Get

Feature	Starter	Plus (Recommended)	Max
M2.7 requests / 5hrs	1,500	4,500	9,000
Image understanding MCP	✅	✅	✅
Web search MCP	✅	✅	✅
Image generation	✅	✅	✅
Speech synthesis	✅	✅	✅
Concurrent OpenClaw agents	—	1-2	3+

Your Plus plan ($200/year) gives you 4,500 M2.7 requests every 5 hours — sufficient for active daily use across all your devices.

Deployment

Backend (Railway / Render / Fly.io)

# Set environment variables
MINIMAX_API_KEY=sk-cp-...
MINIMAX_API_HOST=https://api.minimax.io
JWT_SECRET=<generate-with-openssl-rand-base64-32>

# Deploy
cd backend && npm install && npm run build

Web (Vercel / Cloudflare Pages / Netlify)

cd web && npm install && npm run build
# Deploy the `dist/` folder

Mobile (EAS Build / TestFlight / Google Play)

cd mobile
eas login
eas build --platform ios     # iOS
eas build --platform android # Android

Desktop (GitHub Releases / Homebrew)

cd desktop && npm run tauri build
# Outputs: .exe (Windows), .dmg (macOS), .AppImage (Linux)

Environment Variables Reference

Variable	Required	Description
`MINIMAX_API_KEY`	✅	From platform.minimax.io
`MINIMAX_API_HOST`	✅	`https://api.minimax.io` (global) or `https://api.minimaxi.com` (China)
`JWT_SECRET`	✅	Generate: `openssl rand -base64 32`
`PORT`	No	Default: `3000`
`HOST`	No	Default: `0.0.0.0`

Troubleshooting

"Connection refused" on localhost:3000

The backend isn't running. Start it with cd backend && npm run dev.

Streaming stops after a few seconds

Check your MiniMax API key is valid and the subscription is active at platform.minimax.io.

Rate limit hit immediately

Another device or session is consuming your quota. The 4,500 req/5hr limit is shared across all clients.

Web app CORS errors

Ensure the backend CORS config allows your frontend origin. Update origin: true in backend/src/index.ts to list your domain in production.

Tauri build fails on macOS

Install Xcode Command Line Tools: xcode-select --install

Contributing

Fork the repository
Create a feature branch: git checkout -b feat/your-feature
Commit: git commit -m 'feat: add something'
Push: git push origin feat/your-feature
Open a Pull Request

License

MIT — see LICENSE

Built with MiniMax Token Plan · platform.minimax.io

Name		Name	Last commit message	Last commit date
Latest commit History 24 Commits
backend		backend
desktop		desktop
mobile		mobile
supabase		supabase
web		web
.gitignore		.gitignore
README.md		README.md

Folders and files

Latest commit

History

Repository files navigation

MiniChat — Enterprise AI Chatbot

Features

Screenshots

Architecture

Tech Stack

Frontend

Backend

MiniMax APIs

Getting Started

Prerequisites

1. Clone & Install

2. Configure Environment

3. Start the Backend

4. Start the Web App

5. Run the Mobile App

6. Build the Desktop App

Project Structure

API Reference

Authentication

Conversations

Chat

Rate Limiting

MiniMax Token Plan — What You Get

Deployment

Backend (Railway / Render / Fly.io)

Web (Vercel / Cloudflare Pages / Netlify)

Mobile (EAS Build / TestFlight / Google Play)

Desktop (GitHub Releases / Homebrew)

Environment Variables Reference

Troubleshooting

"Connection refused" on localhost:3000

Streaming stops after a few seconds

Rate limit hit immediately

Web app CORS errors

Tauri build fails on macOS

Contributing

License

About

Topics

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages