Guides
Hunyuan3D for Blender: how to use Tencent's Hunyuan3D add-on inside Blender
Hunyuan3D for Blender: run the Hunyuan3D API server, install Tencent's Blender add-on, generate in the sidebar, plus licence and VRAM notes.
2 min read

Quick answer
Is there a Hunyuan3D Blender add-on?
Yes. Tencent's Hunyuan3D-2 repository includes blender_addon.py, released January 27, 2025. It adds a Hunyuan3D-2 tab to the 3D Viewport sidebar and sends requests to a Hunyuan3D API server you run yourself, by default at http://localhost:8080. Generated models appear in your scene.
How the Hunyuan3D Blender add-on works
Hunyuan3D is Tencent's family of image- and text-to-3D models. The open versions (Hunyuan3D 2.0, 2mini, 2mv and 2.1) can run on your own machine. The Hunyuan3D-2 repository ships blender_addon.py, a Blender add-on that doesn't run the model itself. It's a client:
- You run Hunyuan3D's API server (
api_server.py) somewhere with a capable GPU. - The add-on sends your prompt or image to that server and loads the result into Blender.
The add-on declares Blender 3.0 as its minimum and appears at View3D ‣ Sidebar ‣ Hunyuan3D-2.
Step 1: run the Hunyuan3D API server
Install Hunyuan3D-2 per its README (PyTorch for your platform, then its requirements), then start the server:
python api_server.py --host 0.0.0.0 --port 8080
The README says Hunyuan3D 2.0 needs about 6 GB VRAM for shape generation and 16 GB for shape plus texture. Full install guide: run Hunyuan 3D locally.
Step 2: install the Blender add-on
- Download
blender_addon.pyfrom the repository. - Edit ‣ Preferences ‣ Add-ons ‣ Install from Disk, pick the .py file, enable it.
- Press N in the 3D Viewport and open the Hunyuan3D-2 tab.
Step 3: connect and generate
- In the tab, set API URL. The default is
http://localhost:8080; change it if the server runs on another machine. - Enter a prompt or choose an image and generate.
- The new mesh appears in your scene. If you had a mesh selected, the add-on places the result at that object's location.

Licence: read before you ship
Hunyuan3D's open models use the Tencent Hunyuan community licence, which does not apply in the European Union, the United Kingdom or South Korea. If you're in those regions, you aren't licensed to use the weights. Details: open-source 3D AI licences.
Common problems
- Add-on can't connect: the server isn't running, or the URL/port is wrong. Open the URL in a browser from the Blender machine.
- Out of memory on the server: try the
--low_vram_modeflag shown in the README, or the lighter 2mini model. - Model looks rough: Hunyuan's texture stage needs more VRAM; generate shape only, then texture separately.
Newer Hunyuan models without a GPU
The open add-on targets the Hunyuan3D 2.x models; for the newer 3.x models, check Tencent's repositories for which versions have open weights. Vi3W runs Hunyuan 3D v3.1 Pro, 3.1 Fast, 3.0, 2.1, 2 and 2 Mini in the browser with no GPU, alongside Meshy, Tripo, Rodin and TRELLIS, and exports GLB, FBX or OBJ for Blender. Try Vi3W.
Frequently asked questions
hunyuan3d for blender
Run Hunyuan3D's api_server.py on a machine with a suitable GPU, install blender_addon.py in Blender, and point its API URL at that server. Then generate from the Hunyuan3D-2 tab in the sidebar.
hunyuan 3d to blender
Without the add-on, export GLB from Hunyuan3D's Gradio app or a ComfyUI workflow and import it with File, Import, glTF 2.0. See Hunyuan 3D to Blender for cleanup steps.
Last reviewed

