# codex_api

**Repository Path**: chees_cn/codex_api

## Basic Information

- **Project Name**: codex_api
- **Description**: codex的国内api接口转发服务，适用于服务端部署，然后给本地的codex使用，支持deepseek
- **Primary Language**: Unknown
- **License**: Apache-2.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-06-07
- **Last Updated**: 2026-06-09

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Codex API — Multi-Tenant AI Forwarding Platform

> **Version:** v1.0.1

A multi-tenant AI model forwarding platform built on Node.js + Vue 3, fully compatible with the OpenAI API format (Chat Completions and Responses API). Supports multiple model providers including DeepSeek, Kimi, OpenAI, and SiliconFlow. Can be used as a unified backend proxy for AI coding tools such as Codex CLI, Cursor, Continue, and Aider.

---

## Features

### Core Capabilities
- **Dual API Support**: Both OpenAI Chat Completions API (`/v1/chat/completions`) and Responses API (`/v1/responses`)
- **Multiple Model Providers**: DeepSeek, Kimi, OpenAI, SiliconFlow, Kimi Code — switch freely
- **Multi-Tenant Architecture**: User registration/login, independent API Key management, independent service configuration
- **Streaming Responses**: SSE streaming and non-streaming responses with real-time content
- **Tool Calls**: Auto-fixes various non-standard tool formats, supports multi-tool parallel calling
- **Model Name Mapping**: Client can send any model name; server auto-matches the user's configured provider model
- **Request Logging**: Complete request/response logs with usage statistics and debugging
- **Health Check Endpoint**: `/health` for service monitoring

### Frontend Admin Panel
- **Dashboard**: Usage statistics, request trend charts
- **API Key Management**: Create, view, disable API Keys
- **Provider Configuration**: System presets + user-defined model providers
- **Service Management**: Create forwarding services, bind providers and models
- **Request Logs**: View historical request details

### Kimi Code API Support (v1.0.1)
- **Coding Agent Identity Headers**: Auto-injects `KimiCLI` request headers (`User-Agent`, `X-Msh-Platform`, etc.) to avoid 403 errors
- **Dedicated Endpoint**: Supports `https://api.kimi.com/coding/v1` coding agent API

---

## Project Structure

```
codex-api/
├── backend/                    # Express + Prisma + MySQL backend
│   ├── app.js                  # App entry
│   ├── src/
│   │   ├── config/             # Config (database, env)
│   │   ├── controllers/        # Controllers (Auth, API Key, Provider, etc.)
│   │   ├── middleware/         # Middleware (JWT auth, API Key auth)
│   │   ├── routes/             # Route definitions
│   │   ├── services/           # Business logic (proxy forwarding core)
│   │   └── config/
│   ├── prisma/
│   │   ├── schema.prisma       # Database schema
│   │   └── seed.js             # Seed data (system providers)
│   ├── .env                    # Environment variables (local)
│   └── package.json            # Backend dependencies
│
├── frontend/                   # Vue 3 + Element Plus frontend
│   ├── src/
│   │   ├── views/              # Pages (login, register, dashboard)
│   │   ├── components/         # Common components
│   │   ├── stores/             # Pinia state management
│   │   ├── router/             # Router config
│   │   └── api/                # API client
│   └── package.json            # Frontend dependencies
│
└── README.md                   # This document (Chinese)
└── README_EN.md                # This document (English)
```

---

## Quick Start

### Requirements

- Node.js 18+
- MySQL 8.0+

### 1. Clone

```bash
git clone <repository-url>
cd codex-api
```

### 2. Database Configuration

Create a MySQL database, then configure backend environment variables:

```bash
cd backend
cp .env.example .env
```

Edit `backend/.env`:

```env
# Service port
PORT=8001

# Database configuration (split fields for easier ops)
# Database localhost:3306
DB_HOST=localhost
DB_PORT=3306
DB_USER=username
DB_PASSWORD=password
DB_NAME=codexapi
DB_CONNECTION_LIMIT=10

# JWT config
JWT_SECRET=your-super-secret-jwt-key-change-in-production
JWT_EXPIRES_IN=7d
```

### 3. Initialize Database

```bash
cd backend
npx prisma migrate dev --name init
npx prisma generate
npm run db:seed       # Insert system preset providers (DeepSeek, Kimi, OpenAI, SiliconFlow, Kimi Code)
```

### 4. Start Backend

```bash
cd backend
npm install
npm start
```

Backend will start at `http://127.0.0.1:8001`.

### 5. Start Frontend

```bash
cd frontend
npm install
npm run dev
```

Frontend dev server will start at `http://127.0.0.1:5173`.

---

## System Preset Providers

After running `npm run db:seed`, the system will auto-create the following preset providers:

| Name | Display Name | Base URL | Description |
|------|-------------|----------|-------------|
| `deepseek` | DeepSeek | `https://api.deepseek.com` | DeepSeek AI |
| `kimi` | Moonshot AI (Kimi) | `https://api.moonshot.cn` | Moonshot AI |
| `kimi-code` | Kimi Code | `https://api.kimi.com/coding/v1` | Coding Agent (subscription required) |
| `openai` | OpenAI | `https://api.openai.com` | OpenAI GPT |
| `siliconflow` | SiliconFlow | `https://api.siliconflow.cn` | SiliconFlow |

---

## API Endpoints

### Proxy Endpoints (API Key authentication required)

| Method | Path | Description |
|--------|------|-------------|
| POST | `/v1/chat/completions` | Chat Completions API (streaming/non-streaming) |
| POST | `/v1/responses` | Responses API (streaming/non-streaming) |

### Admin Endpoints

| Method | Path | Description |
|--------|------|-------------|
| POST | `/api/auth/register` | User registration |
| POST | `/api/auth/login` | User login |
| GET | `/api/auth/me` | Get current user info |
| GET | `/api/providers` | Get system provider list |
| GET | `/api/user-providers` | Get user's provider configs |
| POST | `/api/user-providers` | Create user provider config |
| GET | `/api/services` | Get user's service list |
| POST | `/api/services` | Create forwarding service |
| GET | `/api/api-keys` | Get API Key list |
| POST | `/api/api-keys` | Create API Key |
| GET | `/api/request-logs` | Get request logs |
| GET | `/health` | Health check |

### API Key Authentication

Include the API Key in the request header:

```http
Authorization: Bearer sk-your-api-key-here
```

---

## Client Configuration Guide

### Codex CLI

```bash
export OPENAI_API_KEY="sk-your-api-key"
export OPENAI_BASE_URL="http://127.0.0.1:8001/v1"
codex "Hello, world!"
```

### Cursor

In Cursor Settings → Models, add a custom model:
- **Model**: `deepseek-chat` (or any model name you configured)
- **Base URL**: `http://127.0.0.1:8001/v1`
- **API Key**: `sk-your-api-key`

### Continue (VS Code)

Add an OpenAI-compatible provider in `~/.continue/config.json`:

```json
{
  "models": [{
    "title": "Codex API",
    "provider": "openai",
    "model": "deepseek-chat",
    "apiBase": "http://127.0.0.1:8001/v1",
    "apiKey": "sk-your-api-key"
  }]
}
```

### Generic OpenAI SDK

```javascript
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'sk-your-api-key',
  baseURL: 'http://127.0.0.1:8001/v1',
});

const stream = await client.chat.completions.create({
  model: 'deepseek-chat',
  messages: [{ role: 'user', content: 'Hello!' }],
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content || '');
}
```

---

## Request Examples

### Chat Completions API

```bash
curl http://127.0.0.1:8001/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-your-api-key" \
  -d '{
    "model": "deepseek-chat",
    "messages": [
      { "role": "system", "content": "You are a helpful assistant." },
      { "role": "user", "content": "Hello" }
    ],
    "stream": true
  }'
```

### Responses API

```bash
curl http://127.0.0.1:8001/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-your-api-key" \
  -d '{
    "model": "deepseek-chat",
    "input": "Hello, what is 2+2?",
    "stream": true
  }'
```

### Tool Calls

```bash
curl http://127.0.0.1:8001/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-your-api-key" \
  -d '{
    "model": "deepseek-chat",
    "messages": [{ "role": "user", "content": "What is the weather in Beijing?" }],
    "tools": [{
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "Get weather information",
        "parameters": {
          "type": "object",
          "properties": {
            "city": { "type": "string", "description": "City name" }
          },
          "required": ["city"]
        }
      }
    }],
    "stream": false
  }'
```

---

## Version History

### v1.0.1 (Current)
- Added Kimi Code API support with automatic coding agent identity header injection
- Split DATABASE_URL into DB_HOST/PORT/USER/PASSWORD/NAME for easier operations
- Added `env.js` auto-builder to compose DATABASE_URL from split variables
- Updated README with v1.0.1 version and full architecture docs
- Added `kimi-code` preset provider in seed data
- Added Codex CLI tutorial page in frontend

### v1.0.0
- Initial multi-tenant AI forwarding platform
- Support for Chat Completions and Responses API dual formats
- Support for DeepSeek, Kimi, OpenAI, SiliconFlow providers
- Frontend admin panel (Vue 3 + Element Plus)
- User system, API Key management, service configuration, request logs

---

## Tech Stack

### Backend
- **Node.js** — Runtime
- **Express** — HTTP framework
- **Prisma** — ORM database access
- **MySQL** — Data persistence
- **OpenAI Node.js SDK** — Upstream API client (OpenAI-compatible)
- **jsonwebtoken / bcryptjs** — JWT authentication and password hashing
- **dotenv** — Environment variable management

### Frontend
- **Vue 3** — Progressive framework
- **Element Plus** — UI component library
- **Pinia** — State management
- **Vue Router** — Routing
- **ECharts** — Data visualization
- **Vite** — Build tool

---

## License

MIT