1. Create an API key
Create a separate key for every project so usage can be limited and audited later.
Authorization: Bearer YOUR_API_KEYOpenAI API Proxy · OpenAI Compatible API
OpenAI API Proxy
Use one aliapi.me Base URL for OpenAI-compatible requests. Keep the SDK you already use while adding project keys, request logs, usage analytics and future multi-model routing.
Setup
Configure OpenAI API proxy in three steps.
This page targets real developer intent: where to put the Base URL, how to update SDK setup and how to debug failed requests.
2. Replace Base URL
Set your SDK or tool to the aliapi.me OpenAI-compatible endpoint.
https://api.aliapi.me/v13. Keep request shape
Continue using Chat Completions, models and messages in the familiar OpenAI format.
/v1/chat/completionsCapabilities
Why use a gateway instead of scattered API keys?
| Capability | Direct integration | With aliapi.me |
|---|---|---|
| Base URL | Configured per provider | One entry point for future routing |
| Keys | Spread across projects and env vars | Create, disable, limit and audit per project |
| Debugging | Implemented by each app | Central logs for status, latency, tokens and errors |
| Model switching | Requires code or config changes | Can be moved into gateway policy |
FAQ
OpenAI API Proxy FAQ
Do I need to rewrite my OpenAI SDK code?
No. The common path is to keep the SDK and replace only the Base URL and API key.
Can this work with Dify, FastGPT, Cursor and LangChain?
Those tools usually support a custom OpenAI-compatible Base URL, so they can connect to different models through one gateway.
How should the service be described?
Use clear terms such as API relay, unified access, API gateway and OpenAI-compatible API. Avoid unclear or exaggerated claims.
Expand from OpenAI proxy setup to multi-model access.
Browse Claude, Gemini, DeepSeek, Qwen and other model groups in the model directory, or use the FAQ library for Base URL, SDK and tool setup questions.