openai-api-proxy

Simple proxy for OpenAi api via a one-line docker command

简体中文

以下英文由GPT翻译。The following English was translated by GPT.

NodeJS Deployment

You can deploy ./app.js to any environment that supports nodejs 14+, such as cloud functions and edge computing platforms.

1. Copy app.js and package.json to the directory
2. Run yarn install to install dependencies
3. Run node app.js to start the service.

Docker Deployment

docker run -p 9000:9000 easychen/ai.level06.com:latest

The proxy address is http://${IP}:9000.

Available Environment Variables

1. PORT: Service port.
2. PROXY_KEY: Proxy access key used to restrict access.
3. TIMEOUT: Request timeout, default is 5 seconds.

Usage of the API

1. Change the request address of OpenAI (https://api.openai.com) to the address of this proxy (without a slash).
2. If PROXY_KEY is set, add :<PROXY_KEY> after the OpenAI key. If not set, no modification is required.

Explanation

1. Only GET and POST method interfaces are supported, and file-related interfaces are not supported.
2. SSE is currently not supported, so stream-related options need to be turned off. SSE is OK now.

Example of Client Usage

Take https://www.npmjs.com/package/chatgpt as an example.

chatApi= new gpt.ChatGPTAPI({
    apiKey: 'sk.....:<proxy_key here>',
    apiBaseUrl: "http://localhost:9001", // Pass the proxy address
});

Acknowledgments

1. SSE referenced the related code in chatgpt-api project.