Home

Yokai

One terminal to deploy, monitor, and manage LLM inference across all your GPUs.

Yokai is an open-source, terminal-first GPU fleet manager for running vLLM, llama.cpp, and ComfyUI across one or more machines. It combines device onboarding, curated Best-Known-Configs, guided deployment, live fleet monitoring, service lifecycle controls, and AI coding-tool configuration in one cross-platform application.

Official documentation: spencerbull.github.io/yokai-docs

Get started

1. Install Yokai
2. Run the quick start
3. Configure Yokai

Common tasks

• Add GPU devices
• Tag Tailscale GPU compute nodes
• Deploy models
• Use Best-Known-Configs
• Monitor the fleet
• Configure AI coding tools
• Update Yokai

Reference

• CLI commands
• Keyboard shortcuts
• Security model
• Troubleshooting
• System architecture

Project links

• Source code
• Latest release
• Issues and feature requests
• Contributing guide
• MIT license