Skip to main content
Point your coding CLIs, agents, and desktop clients at OmniMux with a Base URL and API key so they use models enabled for your account.

​
Common settings

See Authentication and Models.

​
Coding CLIs

Claude Code CLI

Anthropic official CLI via Anthropic-compatible gateway.

Codex CLI

OpenAI official CLI with an OpenAI-compatible provider.

Gemini CLI

Google Gemini command-line tool.

OpenCode

Open-source coding agent with a custom provider.

Pi Coding Agent

Anthropic Messages-compatible setup.

Hermes Agent

OpenAI-compatible custom endpoint.

​
Desktop and IDE

Claude Desktop

Desktop gateway configuration.

CodeBuddy / WorkBuddy

IDE / plugin-style clients.

​
OpenClaw (chat channel gateway)

Manual install

Install OpenClaw Gateway and configure OmniMux multi-protocol providers.

Auto install

Deploy with OpenClaw Manager.

OpenClaw + Feishu

Feishu / Lark bot setup.

OpenClaw + Telegram

Telegram bot setup.

​
Checklist

  1. Create a token in the console and confirm model scope
  2. Use api.omnimux.ai as the API host — not the console host
  3. Confirm real model IDs via GET /v1/models or the console
  4. Smoke-test with a short chat before wiring agents or channels
Full HTTP surface: API manual.