Ino Get Video Model Download Config

ComfyUI-InoNodes has a strong opinion about video: you shouldn't have to hunt down where a model file lives or remember which Civitai version ID is which. So it ships with a built-in catalog of known video models - and Ino Get Video Model Download Config is the node that turns "I want this one" into a complete download config JSON, ready to feed a downloader. One dropdown, one JSON string out. That's the whole job.

It's one of the pack's InoModelHelper nodes, the family that makes the "download on demand" pipeline work. The philosophy here is worth understanding before you use it: instead of you supplying repo IDs and filenames by hand, you pick from a curated list and the node fills in the details.

The input

Just one required input: model, a dropdown. The options are read at import time from a bundled CSV (video_models_files.csv), and the list is genuinely video-focused: Wan 2.2 diffusion models in fp16 (both text-to-video and image-to-video variants, high and low noise), Civitai-hosted remixes, and the like. You'll see names like comfy-wan-22-T2V-14B-high-fp16 and civit-remix-v21-I2V-low-fp16.

The single output, ModelConfig, is a JSON string containing everything a downloader needs: the host (HF or Civitai), model type, target subfolder, repo ID or Civitai model ID, filename or file ID, and any subfolder/revision info.

How it works

It looks up your chosen name in the CSV, finds the matching row, and serializes it to the same JSON config format the rest of the pack uses. That's why the output plugs directly into Ino Handle Download Model (which reads the host and routes to the right downloader) or the individual HuggingFace/Civitai download nodes.

The catch you need to know

This is the author's catalog, not yours. The dropdown is fixed to what ships in the pack's data/ folder - it won't include your personal video model, and it may lag behind new releases. If the model you want isn't in the list, you're not stuck: Ino Create Model File Config builds a config from scratch, and Ino Http Download Model takes a raw URL. Use this node when the model you want is in the catalog (the fastest path), and reach for the manual builders when it isn't.

Where it fits

The intended flow: Ino Get Video Model Download Config → Ino Handle Download Model → the model lands in models/diffusion_models/<subfolder>/ → a loader node picks it up. The pack's model system is built to download-and-load in one automated chain, so video generation workflows can pull their Wan 2.2 base at runtime instead of requiring it pre-installed. If you're running automated pipelines, that's a real convenience; if you're a one-off hobbyist, it's arguably over-engineered compared to just dropping the file in models/.

Gotchas

Remember the CSV is read when the node loads - if you update the pack's data files you'll need to refresh ComfyUI to see new options. And enabled matters as always: disabled returns an empty string, so feed it to a downloader and you get a confusing failure. Finally, actually downloading a Civitai-hosted model may want a CIVITAI_TOKEN environment variable; HF-hosted ones (like the Comfy-Org Wan repackages) usually don't.

Installation

Ships with ComfyUI-InoNodes:

• ComfyUI Manager: search "ComfyUI Ino Nodes", install, restart.
• Manual: cd ComfyUI/custom_nodes && git clone https://github.com/nobandegani/comfyui_ino_nodes && cd comfyui_ino_nodes && pip install -r requirements.txt, restart.

Python 3.10+, ComfyUI 0.18.1+ (V3 schema). No models download at install - the node's whole purpose is to fetch them at runtime.