📥👑📹 Load Lora Hunyuan Video (+Download from CivitAi)

This is the HunyuanVideo LoRA version of Bjornulf's CivitAI loaders. You pick a Hunyuan Video LoRA from a curated in-node gallery (about 161 entries), and the node downloads it from CivitAI if you don't already have it, then applies it to your Hunyuan model and CLIP. Same shape as the other LoRA selectors in the pack - the difference is these LoRAs are trained for Tencent's HunyuanVideo model, not for image checkpoints.

Quick context so you know what you're getting into: HunyuanVideo was the December 2024 model that made local video generation feel real, and it became a favorite for LoRA training because it trains easily and is uncensored out of the box. That's why there's a big pool of Hunyuan LoRAs on CivitAI to pull from.

How it works

You feed it your model and clip, choose a LoRA, and on run the node checks for the file locally; if it's missing, it downloads it from CivitAI into the pack's Hunyuan LoRA folder and patches it into your model and CLIP at the strengths you set. It's a normal LoRA apply - the file just comes from CivitAI. This node applies the LoRA; you still need the rest of a working HunyuanVideo workflow (the model, the sampler, the decode) around it.

The inputs and outputs that matter

• image - the LoRA picker, a thumbnail gallery of roughly 161 Hunyuan Video LoRAs. Your main choice.
• model and clip - from your HunyuanVideo model loader.
• strength_model and strength_clip - LoRA influence (default 1.0; the −20 to 20 range is wide, useful values sit around 0.5–1.2).
• civitai_token - CivitAI API token, only needed for the download.

Outputs: model and clip to chain into your Hunyuan sampler, plus name, civitai_url, and trigger_words strings. The author includes a ready-made workflows/HUNYUAN_basic_lora.json in the repo if you want a working example to start from.

How to install it

ComfyUI Manager: search Bjornulf_custom_nodes, install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/justUmen/Bjornulf_custom_nodes
pip install -r Bjornulf_custom_nodes/requirements.txt

Restart ComfyUI. Downloads depend on civitai-py from that requirements file. Windows portable: .\python.exe -m pip install ....

Common issues

Same download caveats as the rest of the family: you need a valid token, and gated LoRAs can refuse. Beyond that, the bigger thing to understand is the base model. HunyuanVideo is VRAM-hungry - the original 13B model wanted 45–60GB, and even the newer 8.3B HunyuanVideo 1.5 lands around 10–14GB with offloading. A Hunyuan LoRA is useless without a Hunyuan video pipeline behind it, so make sure you've actually got that set up first.

One honest note on the ecosystem: as of 2026, HunyuanVideo has largely been overtaken by Wan for general local video work, so if you're just starting a video setup, check whether your target LoRAs are Hunyuan or Wan before committing. This node is for Hunyuan LoRAs specifically; the SD1.5, SDXL, and Pony siblings cover the image bases.