Status: [LIVE]

Endpoints found:
https://open.bigmodel.cn/api/paas/v4`
https://ark.cn-beijing.volces.com/api/v3`

Source: https://github.com/760985933/codex_switch

# Nettopo Switch (Codex ↔ LLM Local Proxy)

[中文](README.zh.md)

## Overview

Nettopo Switch is a local proxy that adapts Codex-compatible `POST /v1/responses` (including SSE streaming) and forwards requests to various LLM providers' `POST /v1/chat/completions`. It provides a desktop UI to manage endpoints, API keys, ports, and model mappings across multiple providers, plus tools to manage Codex `config.toml` with safe backups.

This is designed to work with Codex Desktop by pointing Codex's Base URL to the local proxy, so Codex can use different LLM models without changing its workflow.

## Features

- **Codex Responses adapter**: supports `POST /v1/responses` (SSE) and forwards to provider `POST /v1/chat/completions`
- **Multi-provider support**: switch between 9+ built-in LLM providers, or add custom ones
- **Model mapping**: maps Codex-side model names (e.g. `gpt-5.4-mini`) to provider-specific models
- **Menu bar balance display**: shows real-time available usage balance in the system tray (for supported providers like DeepSeek)
- **Visual configuration**: configure Base URL, API key, port, and mappings in the desktop app
- **Codex config.toml management**: merge-write, edit raw content, create/restore/delete/clean backup history
- **Health check & logs**: one-click upstream connectivity check; logs for each request
- **UI i18n**: `zh-CN` (简体中文), `en-US` (English), `ja-JP` (日本語), `ko-KR` (한국어), `fr-FR` (Français), `de-DE` (Deutsch), `es-ES` (Español)
- **Cross-platform builds**: macOS arm64 / Windows amd64 / Windows arm64

## Supported Providers

| Provider | Default Base URL | Default Model |
|---|---|---|
| DeepSeek | `https://api.deepseek.com/v1` | `deepseek-v4-flash` |
| 阿里通义千问 (Alibaba) | `https://dashscope.aliyuncs.com/compatible-mode/v1` | `qwen3.6-plus` |
| 小米 MiMo (Xiaomi) | `https://api.xiaomimimo.com/v1` | `mimo-v2.5-pro` |
| 智谱 GLM (Zhipu) | `https://open.bigmodel.cn/api/paas/v4` | `glm-4.7-flash` |
| �