GPT Proxy

A simple OpenAI API key management platform for efficiently managing and optimizing API key pools, implementing proxy requests and key rotation.

Features

Key Management

• Bulk Key Addition: Add multiple OpenAI API keys at once
• Key Status Management: Maintain valid/invalid status of keys
• Automatic Validation: Automatically validate the validity of keys
• Batch Reset: Reset all invalid keys to valid status with one click

Request Proxy and Load Balancing

• Intelligent Proxy: Automatically route requests to valid API keys
• Load Balancing: Evenly distribute requests to optimize key usage

Data Statistics and Monitoring

• Real-time Statistics: Display call statistics for the last 1 minute, 1 hour, 24 hours, and total
• Key Usage Tracking: Record the number of uses and last used time for each key
• Key Pool Status: Display valid/invalid key count statistics

User Interface

• Responsive Design: Adapt to different device screen sizes
• Paginated Browsing: Efficiently browse and manage large numbers of keys
• Intuitive Operation: Clear and straightforward operation interface

Installation and Setup

Requirements

• Python 3.8+
• FastAPI
• SQLAlchemy
• Other dependencies (see requirements.txt)

Installation Steps

1. Clone the repository

   git clone https://github.com/Inblac/gpt-proxy.git
   cd gpt-proxy

2. Create a virtual environment and install dependencies

   python -m venv venv
   source venv/bin/activate  # Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Create a data directory and configure config.ini

   mkdir data
   cp config.ini.example data/config.ini
   # Edit data/config.ini as needed

4. Initialize the database (will be created as ./data/gpt_proxy.db)

   python -m gpt_proxy.database.init_db
   # This will create data/gpt_proxy.db if it does not exist

5. Start the service

   uvicorn gpt_proxy.main:app --host 0.0.0.0 --port 8000

Docker & Docker Compose

• Both config.ini and gpt_proxy.db must be placed in the ./data directory.
• When using Docker or docker-compose, mount the host's ./data directory to the container's /data directory:

docker-compose.yml example:

version: '3.8'
services:
  gpt-proxy:
    build: .
    ports:
      - "8000:8000"
    volumes:
      - ./data:/data
    restart: unless-stopped

This ensures all configuration and data are persistent and easy to manage.

Usage Instructions

Admin Login

1. Visit the system homepage http://localhost:8000/
2. Log in using the admin API Key

Adding API Keys

1. In the admin interface, find the "Add New OpenAI API Key" section
2. Enter one or more API keys (one per line)
3. Click the "Add Keys" button

Managing Key Status

• Click "Set Invalid" or "Set Valid" in the key list to toggle key status
• Use the "Revalidate All Invalid Keys" button to verify key validity
• Use the "Reset All Invalid Keys to Valid Status" feature to batch restore keys

Viewing Statistics

• The dashboard displays call statistics and key pool status
• Statistics automatically refresh every 30 seconds

API Interface Description

Proxy Requests

• POST /v1/* - Proxy all requests to the OpenAI API

Management Interfaces

• POST /token - Admin authentication
• GET /api/stats - Get statistics data
• GET /api/keys/paginated - Get paginated key list
• POST /api/keys/bulk - Bulk add keys
• DELETE /api/keys/{key_id} - Delete specified key
• PUT /api/keys/{key_id}/status - Update key status
• POST /api/validate_keys - Trigger key validation

Security Considerations

• Use a strong password as the admin API Key
• Regularly rotate the admin API Key
• Run the system in a protected network environment
• Avoid exposing the admin interface in public environments

Contributing

Bug reports and feature requests are welcome. If you want to contribute code, please open an issue to discuss your ideas first, then submit a PR.

License

MIT License

数据库配置

GPT-Proxy默认使用SQLite数据库存储API密钥和相关数据。现在也支持使用PostgreSQL作为替代选项。

配置数据库类型

在data/config.ini文件中，添加或修改[Database]部分来配置数据库：

[Database]
# 数据库类型: sqlite (默认) 或 postgresql
type = sqlite

# 以下配置仅在 type = postgresql 时使用
host = localhost
port = 5432
database = gpt_proxy
user = postgres
password = your_password

PostgreSQL支持

如果要使用PostgreSQL，需要安装psycopg2库：

pip install psycopg2-binary

然后修改配置文件，将type设置为postgresql，并配置正确的连接参数。