Hunyuan Config

What it is

Hunyuan Config is the credentials node for the whole ComfyUI-Hunyuan3D-v3 pack. It doesn't generate anything itself - it just packages your Tencent Cloud API keys into a HUNYUAN_CONFIG object and hands it downstream to whichever node actually does the work: Hunyuan Image to 3D, Hunyuan Text to 3D, or Hunyuan Batch Image to 3D. Every workflow in this pack starts here, the same way an image workflow starts with a checkpoint loader.

Worth saying up front, because it trips people up: this pack is a thin wrapper around Tencent Cloud's Hunyuan 3D Global API v3, a paid cloud service. It is not the same thing as Tencent's open-weight Hunyuan3D 2.0/2.1 models that people self-host and run on their own GPU with a GGUF quant. Same brand, completely different product - one is a checkpoint you download, the other is a metered API call you pay Tencent for per request. If you were expecting to point this at local weights, you're in the wrong node; if you want a hosted API and don't mind paying per generation, this is exactly that.

How it works

The node takes your secret_id and secret_key - the pair Tencent Cloud issues for programmatic API access - plus a region, and bundles them into the custom HUNYUAN_CONFIG type that the generator nodes expect as their config input. Nothing gets validated or called here; it's pure plumbing. The actual HTTP request to Tencent's API happens in the downstream node when you queue the prompt.

Inputs and outputs

• secret_id / secret_key - plain string fields, the credential pair from your Tencent Cloud account. Treat secret_key like a password: don't share a workflow JSON with these baked in, and don't paste it into a screenshot for a support forum.
• region - currently locked to ap-singapore, matching the "Global API" the README references. There's nothing to configure here beyond leaving the default.

Output is a single config (HUNYUAN_CONFIG) - wire it into the config input of Hunyuan Image to 3D, Hunyuan Text to 3D, or Hunyuan Batch Image to 3D. It's not consumed by the preview nodes; those just take a file path.

Getting the credentials

Before you touch this node, you need a Tencent Cloud account with billing enabled. The README lays out the sequence: create the keys under Access Management → API Keys in the Tencent Cloud console, activate the Hunyuan 3D service under AI Services, then add payment info. Budget for it - this is a paid API at roughly $0.10–0.60 per generation, and Tencent won't let the calls through without credits on the account (the README's own recommendation is $10+ to start).

Fair warning on the console itself: this is the part people actually get stuck on, not the ComfyUI side. It's an international Tencent Cloud account, the documentation is thin in English, and error codes don't always match what's in their own reference - one user reported grinding through trial and error and eventually having to file a support ticket to get an undocumented error resolved (support did answer within the hour, for what it's worth). So if setup feels harder than it should, that's apparently normal, not you doing something wrong.

Install

Via ComfyUI Manager: search "Hunyuan 3D", install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/exedesign/Hunyuan-3D-v3.git
cd Hunyuan-3D-v3
pip install -r requirements.txt

Restart ComfyUI after either method and HunyuanConfig should show up under the Hunyuan3D/v3 category.

Troubleshooting

If a downstream node throws a ResourceInsufficient error, it traces back to this node's credentials, not a bug in the generator - it means the Tencent account behind those keys is out of balance or the Hunyuan 3D service was never actually activated. Check the console balance, confirm the service shows as active under AI Services, and top up if needed.