GitHub radar
Turn one photo into animated 3D hair for your character
HairDeck builds animation-ready hair from a single image for a 3D avatar — it sways when the character moves instead of sitting stiff like a helmet.
It's an AI tool: from one photo it creates hair strips and digital 'bones' that bend them — giving you a hairstyle that moves naturally.
When it helps
Useful if you're building a 3D avatar or VTuber character and want ready-to-animate hair without manual rigging work. Skip it if you don't need a 3D character model or only work with 2D art.
Pros
- Turns one photo into animation-ready hair
- Exports straight to VRM for VTubers and games
- Free online demo needs no GPU of your own
- Saves hours of manual strand placement in Blender
Cons
- Running it locally needs an NVIDIA GPU and Linux
- Won't run on an ordinary laptop without the demo
How to set it up — step by step
- 1Open the HairDeck demo on Hugging Face Spaces at https://huggingface.co/spaces/hyz317/HairDeck — no install needed.
- 2Upload a photo of the character whose hair you want to animate.
- 3Run the generation and review a few hairstyle variations with animation.
- 4Export the result as a VRM avatar or a GLB file for your own project.
- 5Send the agent the text in the block below if you want it to run the demo and explain the result for you.
- 6If you have an NVIDIA GPU and Linux, ask the agent to help set up HairDeck locally using the repo's instructions.
Text for your agent
Copy this and send it to your agent — Claude Code, Codex, any of them:
Open the HairDeck demo on Hugging Face (https://huggingface.co/spaces/hyz317/HairDeck) and generate animated hair from my character photo — tell me what you got and how to export it as VRM. If I have an NVIDIA GPU and Linux, ask whether I want to set up HairDeck locally following https://github.com/hyz317/HairDeck.
You need your own NVIDIA GPU and Linux — this will not run on a regular laptop. But the authors host a free demo on Hugging Face Spaces, so you can try it without your own hardware.
Open on GitHub▌ More finds