Look up which text encoder and VAE a specific image generation model needs before downloading it.
Check whether a VAE or LoRA from one model family will work with a different diffusion model.
Browse a visual family tree to see how finetuned models descend from their base models.
Contribute a correction or a new model entry by editing the project's YAML data files.
| jfcantu/model-codex | 0xkinno/neuralvault | 0xmayurrr/ai-contractauditor | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | easy | hard | easy |
| Complexity | 2/5 | 4/5 | 2/5 |
| Audience | vibe coder | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
No backend or database is needed, npm install and npm run dev start a local static site.
Model Codex is a browsable reference site for people who use ComfyUI, a tool for running AI image generation models. Image models like Flux or Stable Diffusion XL often need extra companion files to actually work, such as text encoders that turn a written prompt into something the model understands, and VAEs that turn the model's internal output into a final image. Figuring out which companion files go with which model, and whether two models are compatible with each other, can be confusing, and this project exists to answer exactly that question in plain language. The site itself is static, meaning it has no backend server or database. All the actual knowledge lives in hand editable YAML files, and a React and Vite based front end reads those files and renders them as pages. There is a Quickstart page for common tasks, a Browse page split into diffusion models, text encoders, and VAEs, a Model Tree page showing how models descend from one another, a Compatibility checker where you pick a VAE, a model family, and a diffusion model and get a plain language answer of yes, probably, maybe, or no, and a Glossary explaining terms used throughout the site. Under the hood, every model is treated as a node in a family tree, and a model that does not declare its own required files simply inherits them from its nearest ancestor. Compatibility exceptions and additions are recorded separately, since the checker already assumes sensible defaults based on a model's family and lineage. The project's author is upfront that this was built quickly with AI assistance and that both the code and the compatibility data should be treated as a helpful starting point rather than a guaranteed source of truth, with unverified entries marked as such. This project is aimed at people already using ComfyUI who want a quick, understandable way to check what a model needs or whether two files will work together. The application code is released under the MIT license, while the curated data itself is placed in the public domain under CC0, so both can be reused freely without needing permission.
A browsable reference site that explains which text encoders, VAEs, and image models are compatible with each other in ComfyUI.
Mainly TypeScript. The stack also includes TypeScript, React, Vite.
The app code is free to use, modify, and distribute for any purpose including commercial use, as long as the copyright notice is kept, the data is dedicated to the public domain with no restrictions at all.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
double-check against the repo, no cap.