Gerbil Studio

Plan a fully local LLM setup

Gerbil Studio — plan a fully local LLM setup on your own hardware with KoboldCpp models, GPU options, and image workflow

How to use Gerbil Studio

  1. Pick your operating system, GPU acceleration, and model size tier
  2. Read the RAM estimate and the CPU versus GPU advice for your chosen model size
  3. Follow the five setup steps, from the release download to your first offline chat
  4. Copy the plan and pick an image generation preset: Flux, Chroma, Qwen Image, or Z-Image
  5. Download the real app from the Gerbil releases page and import your own GGUF model

Features

  • Built on the real lone-cloud gerbil repo, version 1.29.1, AGPL-3.0 licensed
  • Powered by KoboldCpp, a llama.cpp fork, with CPU, CUDA, ROCm, Vulkan, and Metal support
  • Real platform files from the README: Windows portable and setup, macOS dmg, Linux AppImage
  • Real image presets from the README: Flux, Chroma, Qwen Image, and Z-Image workflows
  • RAM estimator based on model size plus 2GB overhead, with CPU slowness warnings
  • Honest scope: no model runs in this page; the real app runs offline on your own machine

Frequently asked questions

Is it free?

Yes, unlimited.

Login to use this free online tool. Create a free account to get started.