chat2api Documentation

You can manage a pool of tokens for multi-account rotation and polling:

1. Set Authorization Key: Configure the AUTHORIZATION environment variable with one or more authorization codes (comma-separated). This acts as your personal API key.
2. Upload Tokens: Access the /tokens or /{api_prefix}/tokens endpoint to upload new AccessToken or RefreshToken values, view the current count, or clear the pool.
3. Requesting: When making API calls, pass the value configured in AUTHORIZATION as the Bearer token in the Authorization header. The system will automatically rotate through the available tokens on request failure.

Choose one of the following deployment methods:

Direct Deployment (Python):

git clone https://github.com/LanQian528/chat2api
cd chat2api
pip install -r requirements.txt
python app.py

Docker Deployment:

docker run -d \
  --name chat2api \
  -p 5005:5005 \
  lanqian528/chat2api:latest

Docker Compose (Recommended for PLUS accounts):

1. Create a directory and enter it.
2. Download the compose file: wget https://raw.githubusercontent.com/LanQian528/chat2api/main/docker-compose-warp.yml
3. Edit the file to set your environment variables.
4. Run: docker-compose up -d

To run a mirror of the official ChatGPT web interface:

1. Enable Gateway: Set the environment variable ENABLE_GATEWAY to true.
2. Upload Tokens: Use the /tokens management page to upload RefreshToken or AccessToken values.
3. Access Login: Navigate to the /login page to access the mirror interface.
4. Direct Login: You can also log in directly using the URL format /?token=xxx, where xxx is a RefreshToken, AccessToken, or SeedToken (random seed).

You can deploy the chat2api service using Docker Compose. The setup includes the main application and a watchtower service to automatically update the container.

By default, the application maps port 5005 on the host to port 5005 in the container. Data is persisted in a local ./data directory.

version: '3'

services:
  chat2api:
    image: lanqian528/chat2api:latest
    container_name: chat2api
    restart: unless-stopped
    ports:
      - '5005:5005'
    volumes:
      - ./data:/app/data
    environment:
      - TZ=Asia/Shanghai
      - AUTHORIZATION=sk-xxx

watchtower:
    image: containrrr/watchtower
    container_name: watchtower
    restart: unless-stopped
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
    command: --cleanup --interval 300 chat2api

If you encounter errors while using the API, check the following codes:

• 401: IP does not support login-free access (try using a US IP or setting PROXY_URL) or authentication failed.
• 403: Check the application logs for specific error details.
• 429: Rate limit exceeded for the current IP. Wait or change IP.
• 500: Internal server error; the request failed.
• 502: Gateway error or network unavailability.

When running chat2api via Docker, you can configure the following environment variables:

• TZ: Sets the timezone (e.g., Asia/Shanghai).
• AUTHORIZATION: Sets the API key/token for authorization (e.g., sk-xxx).

The application behavior changes based on the enable_gateway configuration setting:

• When enable_gateway is True: The application imports and serves various gateway modules (e.g., gateway.share, gateway.login, gateway.chatgpt, etc.).
• When enable_gateway is False: The application enters a restricted mode where any request to a path not explicitly defined will return a 404 Not Found error with the detail "Gateway is disabled".

The Client class provides an asynchronous interface for making HTTP requests using curl_cffi.requests.AsyncSession. It supports proxy configuration, custom timeouts, SSL verification, and TLS impersonation to mimic specific browser fingerprints.

Parameters

• proxy (str, optional): A proxy URL for both HTTP and HTTPS traffic.
• timeout (int, optional): Request timeout in seconds. Defaults to 15.
• verify (bool, optional): Whether to verify SSL certificates. Defaults to True.
• impersonate (str, optional): The TLS fingerprint to impersonate (e.g., 'safari15_3'). Defaults to 'safari15_3'.