Goethe
Turns a Linux GPU machine into your own OpenAI-, Anthropic- and LiteLLM-compatible model API.
- AI tool
- Free
Pricing
Free and open source (Apache-2.0); runs on your own NVIDIA GPU hardware.
Overview
Goethe turns a Linux machine with a supported NVIDIA GPU into a model-serving API that OpenAI, Anthropic and LiteLLM clients can call at one address. You install a model, serve it and choose who can reach it: this machine only, your VPN or a public address, with one key per app or person. Prompts, data and models stay on hardware you control. It installs with a curl script, so read the script before you run it, and keep the endpoint off the public internet unless you need it there.
Key features
- OpenAI, Anthropic and LiteLLM compatible API
- One address for every loaded model
- Per-app API keys
- Local, VPN or public exposure
Use cases
Show all 7Show fewer
Alternatives
Last updated 2026-11-07