Analysis updated 2026-05-18
Run several different AI coding assistants against one local proxy without reconfiguring each client.
Switch between Grok, Claude, Gemini, and OpenAI backends without changing your tool's API settings.
Keep multiple provider accounts' API keys masked when listing configured auths.
| scvxzf1/cliproxyapi-local | ethanbrooks-data/ethanbrooks-data | vaalacat/ai-gateway | |
|---|---|---|---|
| Stars | 10 | 10 | 10 |
| Language | Go | Go | Go |
| Setup difficulty | moderate | easy | moderate |
| Complexity | 4/5 | 1/5 | 4/5 |
| Audience | developer | general | ops devops |
Figures from each repo's GitHub metadata at analysis time.
Requires configuring OAuth or API keys for one or more AI providers before use.
CLIProxyAPI-local is a proxy server that lets you talk to several different AI providers through one consistent interface. It exposes API endpoints that look like OpenAI, Gemini, Claude, Codex, and Grok, so any tool or client already built to speak one of those APIs can be pointed at this proxy instead. The idea is to run it on your own machine and connect multiple accounts across providers such as Kimi, OpenAI, Anthropic Claude, Google Gemini, and xAI Grok, then switch between them without changing how your client software talks to the API. This is aimed at people who already use command line coding tools and want to try different underlying models without rewriting their setup each time. This particular repository is a local fork of the original CLIProxyAPI project, described by its maintainer as adding a few specific changes: routing for xAI Grok's free OAuth signup flow, an embedded usage tracking feature, and masking of secret values when the list of configured accounts is displayed, so tokens are not shown in plain text on screen. Because it runs locally, the proxy sits between your CLI tool and the real provider APIs, forwarding requests and normalizing responses so the client does not need to know which backend is actually answering. It supports OAuth based logins for providers that offer them, alongside standard API key configuration for the rest. The README documents the proxy in English, with additional translations for Chinese and Japanese readers. It is written in Go, and the project is intended for developers who run multiple AI coding assistants side by side and want a single local gateway to manage them, rather than juggling separate configuration for every tool.
A local proxy server that lets OpenAI, Gemini, and Claude-compatible tools talk to Kimi, OpenAI, Anthropic, Google, and xAI Grok through one address.
Mainly Go. The stack also includes Go.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.