OpenAI API Inspect Proxy

A simple proxy server designed to inspect and forward requests to OpenAI-compatible APIs. Useful for debugging, logging, or routing API calls.

---

OpenAI API 检查代理

一个简单的代理服务器，用于检查和转发对 OpenAI 兼容 API 的请求。适用于调试、记录日志或路由 API 调用。

---

Key Features / 主要特性

• Transparent Proxying: Forwards requests to any OpenAI-compatible API endpoint without altering core content.
  • 透明代理: 将请求转发到任何 OpenAI 兼容的 API 端点，不修改核心内容。
• Streaming Support: Handles both standard JSON responses and text/event-stream (Server-Sent Events) for streaming completions.
  • 流式支持: 同时处理标准 JSON 响应和用于流式完成的 text/event-stream (服务器发送事件)。
• Configurable: Target API URL, proxy host, and port can be configured via command-line arguments or environment variables.
  • 可配置: 目标 API URL、代理监听主机和端口可通过命令行参数或环境变量进行配置。
• Request/Response Logging: Logs details of incoming requests and outgoing responses (headers, body) for inspection. The Authorization header is automatically redacted.
  • 请求/响应日志: 记录传入请求和传出响应的详细信息（请求头、请求体）以供检查。Authorization 请求头会自动脱敏处理。
• Stateless API Key Handling: Does not store, manage, or inject API keys. Relies entirely on the Authorization header provided by the client.
  • 无状态 API 密钥处理: 不存储、管理或注入 API 密钥。完全依赖客户端提供的 Authorization 请求头。

---

Usage / 使用方法

Option 1: Using npx (Recommended for quick use)

The package is published to npm:

该包已发布到 npm

npx openai-api-inspect [options]

This command executes the package directly without needing a global installation.

此命令直接执行包，无需全局安装。

Option 2: Using node (After building)

Run the proxy server from the project root directory after building the project (npm run build):

在构建项目 (npm run build) 后，从项目根目录运行代理服务器：

node dist/index.js [options]

Configuration Options / 配置选项:

Options can be provided via command-line flags or environment variables. Command-line flags take precedence.

配置项可以通过命令行标志或环境变量提供。命令行标志优先。

Option	CLI Flag	Environment Variable	Default Value	Description	说明
Target API Base URL	-t, --targetBaseUrl	TARGET_BASE_URL	https://api.ppinfra.com/v3/openai	The base URL of the target OpenAI-compatible API.	目标 OpenAI 兼容 API 的基础 URL。
Proxy Host	-H, --host	PROXY_HOST	0.0.0.0	Host for the proxy server to listen on.	代理服务器监听的主机地址。
Proxy Port	-p, --port	PROXY_PORT	7891	Port for the proxy server to listen on.	代理服务器监听的端口。
Show Help	-h, --help	-	-	Display help information.	显示帮助信息。

Examples / 示例:

• Specify a different target API and port using npx:
  • 使用 npx 指定不同的目标 API 和端口:

  npx openai-api-inspect --targetBaseUrl https://api.openai.com/v1 --port 8080
  # or using aliases
  npx openai-api-inspect -t https://api.openai.com/v1 -p 8080

• Specify a different target API and port using node:
  • 使用 node 指定不同的目标 API 和端口:

  node dist/index.js -t https://api.openai.com/v1 -p 8080

• Use environment variables (works with both npx and node):
  • 使用环境变量 (适用于 npx 和 node):

  export TARGET_BASE_URL=https://api.custom-ai.com/
  export PROXY_PORT=9000
  npx openai-api-inspect 
  # or
  # node dist/index.js

Accessing the Web UI / 访问 Web UI:

Once the proxy is running, you can access the live request/response log viewer in your browser at:

代理运行后，您可以在浏览器中访问实时请求/响应日志查看器：

http://localhost:[PROXY_PORT]/ui/

Replace [PROXY_PORT] with the actual port the proxy is listening on (default is 7891).

将 [PROXY_PORT] 替换为代理正在监听的实际端口（默认为 7891）。

---

Installation / 安装

1. Clone the repository:

   git clone https://github.com/zerob13/openai-api-inspect.git
   cd openai-api-inspect

2. Install dependencies:

   npm install 

3. Build the project (compiles TypeScript to JavaScript):

   npm run build

   (Note: The prepare script usually runs this automatically after npm install)

---

Important Notes / 重要提示

• API Keys: This proxy is transparent regarding authentication. It does not add, remove, or validate API keys. The client application making requests to this proxy must include the correct Authorization: Bearer YOUR_API_KEY header required by the target API.
  • API 密钥: 此代理在身份验证方面是透明的。它不会添加、删除或验证 API 密钥。向此代理发出请求的客户端应用程序必须包含目标 API 所需的正确 Authorization: Bearer YOUR_API_KEY 请求头。
• Target URL: Ensure the targetBaseUrl points to the base path of the target API (e.g., https://api.openai.com/v1), not a specific endpoint like /chat/completions. The proxy forwards the path and query parameters from the original request.
  • 目标 URL: 确保 targetBaseUrl 指向目标 API 的基础路径（例如 https://api.openai.com/v1），而不是像 /chat/completions 这样的具体端点。代理会转发原始请求中的路径和查询参数。
• Logging: Request and response bodies are logged for debugging purposes. Ensure your logging environment is secure if sensitive data might be present in the bodies. API keys in the Authorization header are redacted automatically.
  • 日志记录: 为了调试目的，会记录请求和响应体。如果请求体中可能包含敏感数据，请确保您的日志记录环境是安全的。Authorization 头中的 API 密钥会自动进行脱敏处理。
• HTTPS: Currently, the proxy listens on HTTP. If you need HTTPS for the proxy itself, consider placing it behind a dedicated reverse proxy like Nginx or Caddy that handles TLS termination.
  • HTTPS: 当前，代理服务器监听 HTTP。如果您需要代理服务器本身支持 HTTPS，可以考虑将其部署在专门的反向代理（如 Nginx 或 Caddy）之后，由它们来处理 TLS 终止。

---

Future Plans / 未来计划

• Improve Log Formatting: Enhance the console log output for better readability (e.g., using pino-pretty, adding colors, better structuring).
  • 改进日志格式: 优化控制台日志输出以提高可读性（例如，使用 pino-pretty、添加颜色、优化结构）。
• Enhance JSON Logging: Improve logging of JSON request/response bodies (e.g., pretty-printing, potential truncation for very large bodies).
  • 增强 JSON 日志记录: 改进 JSON 请求/响应体的日志记录（例如，美化打印，对非常大的请求体进行可能的截断）。

---

License / 许可证

MIT