Get started

Overview

OpenLLM routes Claude Code, Codex and your other tools to your subscriptions, API keys, local models and included models through one key, with fallback.

OpenLLM is a routing layer for AI models. You connect the models you have access to, and your tools call all of them through one OpenAI- and Anthropic-compatible endpoint with one sk-llm-… key. When a model is rate-limited or unavailable, OpenLLM sends the request to the next model in your chain.

The parts

  • Hosted gateway at https://www.openllm.sh/v1. It serves API-key and included models, and runs the dashboard where you manage providers, keys, and model chains.
  • Local daemon on your Mac or Linux machine. It serves your AI subscriptions and local model servers, and listens on http://127.0.0.1:8787/v1 by default.
  • CLI and MCP server. The openllm command installs the daemon, starts clients such as Claude Code with OpenLLM applied, and runs an MCP server for your clients.
  • Mac app. Installs and connects the daemon on a Mac without a terminal.
One key in front; the gateway and the daemon each serve the sources they can reach.

Where your models come from

  • Subscriptions. Claude Code, Codex (ChatGPT), Kimi Code, and Grok, signed in through each vendor's own CLI on your device. See Subscriptions.
  • API keys. Anthropic, OpenAI, Google AI, AWS Bedrock, and more, stored encrypted in your vault. See API keys.
  • Local models. Ollama, LM Studio, and other model servers the daemon finds on your machine. See Local models.
  • Included models. Every account includes a monthly allowance on hosted models that OpenLLM serves on its own key, so your first request works before you connect anything. See Included models.

Who it's for

OpenLLM is for developers who use more than one AI tool or model. You run coding agents every day, you pay for an AI subscription or hold API keys from several providers, or you run models locally, and you want all of it behind one key with fallback.

Quickstart

Sign in, copy your key, and make your first request.

How it works

The request path, what runs on your device, and what OpenLLM sees.

Clients

Claude Code, Codex, OpenCode, Raycast, and any OpenAI or Anthropic client.

Providers

Connect subscriptions, API keys, local models, and custom endpoints.

Last updated on

Quickstart

Create an OpenLLM account, set up your vault, copy your key, send your first API request on the included models, and run Claude Code through OpenLLM.

On this page

The partsWhere your models come fromWho it's for