One-click Colabs for Stable Diffusion, now officially retired

What it does The repository is a curated index of Jupyter notebooks that install and launch the popular Stable Diffusion WebUI directly inside Google Colab. Each row in the giant matrix offers three stability tiers—lite, stable, and nightly—pointing at different model checkpoints, from base Stable Diffusion to niche fine-tunes like Waifu Diffusion and Arcane Diffusion. It also branches into DreamBooth/LoRA training and persistent Google Drive installs.

The interesting bit At its peak, this was essentially the fastest way to go from “I want to try anime AI art” to an actual image grid without owning a GPU or typing a shell command. The author did the boring but valuable work of keeping dependency versions, extension lists, and model download URLs in sync so strangers could click a single button and borrow Google’s GPUs for an hour.

Key highlights

• Dozens of model variants (Anything v3, Midjourney v4 diffusion, Cyberpunk Anime, etc.) each with a dedicated notebook.
• Three release channels: lite for minimal extensions, stable for ControlNet v1.1, and nightly for bleeding-edge PyTorch 2.0 builds.
• Separate branches for DreamBooth/LoRA training and for installing the stack to Google Drive.
• 15,940 stars, suggesting it was a default bookmark for the “free Colab GPU” crowd.

Caveats

• The README explicitly states the repository is outdated and no longer receives updates; the author now points users to a successor project (TostUI).
• Because it is unmaintained, the pinned dependency versions and Colab environment hacks are likely rotting.

Verdict Worth a look if you are studying how the early Stable Diffusion community bypassed hardware barriers, or if you need a template for wrapping Python web UIs in Colab. Skip it if you actually want to run inference today—use the author’s newer project or run WebUI locally instead.