Open-Source AI Model Landscape

This repo has moved. Active development now happens at phoenixperry/heldincommon, published at https://heldincommon.org. This copy is kept for history and is no longer updated.

A static reference site cataloguing open-source image, video, and audio generation models — covering license type, training data provenance, attribution mechanisms, commercial-use terms, and known data/IP gaps for each model.

Built for the Federated Data Commons for Creative Communities (FDCCC) project at NYU.

Site

Plain HTML/CSS/JS, no build step or framework:

• index.html — page structure
• styles1.css — styling
• script.js — search, filtering, sorting, and modal logic
• data.js — the model dataset, as a MODELS array

Adding a model

Append an object to MODELS in data.js following the existing field structure, then commit and push — GitHub Pages picks up the change automatically.

All fields:

{
  model: "Model Name",
  developer: "Developer",
  year: "2024",
  modality: "Image" | "Video" | "Audio",
  io: "Text → Image",
  licenseType: "Fully Open" | "Permissive" | "Restricted-Open" | "Mixed",
  licenseName: "Apache 2.0",
  dataDisclosed: "Full" | "Partial" | "No",
  dataSources: "...",
  attribution: "...",
  commercialUse: "Full" | "Yes" | "Partial" | "No",
  finetune: "Yes" | "Partial" | "No",
  checkpoint: "Yes" | "No",
  api: "Yes" | "No",
  exportFormat: "PNG, JPG",
  openSource: "Yes" | "No",
  dataGap: "...",
  notes: "..."
}

Adding a ComfyUI workflow / output

To attach a workflow and/or output sample to a model's detail view, place the files in a subfolder under Image Audio Video Models/:

Image Audio Video Models/
  Your Model Name/
    workflow.json              ← ComfyUI workflow export (optional)
    output.png / .wav / .mp4  ← example output (optional)

Then add a workflow field to the model's entry in data.js. All sub-fields are optional — include only what you have:

workflow: {
  folder: "Your Model Name",   // must match folder name exactly
  json:   "workflow.json",     // ComfyUI workflow — shows Workflow JSON tab
  image:  "output.png",        // image output   ─┐
  audio:  "output.wav",        // audio output    ├ shows Output tab (use one)
  video:  "output.mp4"         // video output   ─┘
}

The modal will show an Output tab (image viewer, audio player, or video player) and/or a Workflow JSON tab depending on what fields are present. Models with no workflow field show only the Details tab.

Note: if a model folder contains a Python virtual environment (.venv/), it is gitignored automatically — do not commit it.

Running locally

Any static file server works, e.g.:

python -m http.server 8080

Then open http://localhost:8080.

Hosting

Served via GitHub Pages from this repo.