[LIVE] [yhfgyyf/codex-api-server] 1 endpoint(s)
ChatGPT
API Proxy/ChatGPT
2,120 characters
Status: [LIVE]
Endpoints found:
https://chatgpt.com/backend-api`
Source: https://github.com/yhfgyyf/codex-api-server
# Codex API Server
[中文文档](README_CN.md)
OpenAI API-compatible proxy server that uses ChatGPT Plus (Codex) OAuth credentials to provide standard `/v1/chat/completions`, `/v1/responses`, and Anthropic `/v1/messages` endpoints.
## Features
- **OpenAI API compatible** — works with any OpenAI SDK, vLLM client, or tool
- **Anthropic Messages API** — full `/v1/messages` support with `thinking` content blocks
- **Reasoning model support** — streams `reasoning_content` (thinking process) in DeepSeek/vLLM format
- **Function calling** — full tool/function calling support with streaming tool_calls
- **Auto token refresh** — reads `~/.codex/auth.json`, refreshes OAuth tokens before expiry
- **Streaming SSE** — full Server-Sent Events support for all endpoints
- **Responses API** — proxy passthrough for OpenAI's `/v1/responses` endpoint
- **vLLM parameter compatible** — accepts all OpenAI/vLLM params, silently ignores unsupported ones
- **Request logging** — all API calls stored in SQLite with full request/response data
- **JSONL export** — download or export logs for analysis and fine-tuning
- **Async & connection pooling** — built on FastAPI + httpx for high concurrency
## Supported Models
- `gpt-5.5`
- `gpt-5.4`
- `gpt-5.4-mini`
- `gpt-5.3-codex`
- `gpt-5.3-codex-spark`
- `gpt-image-2`
## Prerequisites
- Python 3.11+
- A valid `~/.codex/auth.json` file with ChatGPT OAuth credentials (from [OpenAI Codex CLI](https://github.com/openai/codex))
## Installation
```bash
# Create conda environment
conda create -n codex-api-server python=3.11 -y
conda activate codex-api-server
# Install dependencies
pip install -r requirements.txt
```
## Quick Start
```bash
conda activate codex-api-server
python server.py
```
The server starts on `http://127.0.0.1:18888` by default.
## Usage
### With curl
```bash
# List models
curl http://localhost:18888/v1/models
# Chat completion
curl http://localhost:18888/v1/chat/completions \
-H "Content-Type: ap