ComfyUI NSFW on RunPod
This guide runs an ahegao.red pack on a rented RunPod GPU for a few dollars an hour. Your laptop stays the controller, and the studio-class card does the rendering. Models live on a RunPod network volume, so they survive every pod restart. Your prompts and output stay on the pod and the volume.
Prepare your RunPod account
- Create an account at runpod.io and add billing credits under Billing.
- Under Storage, create a Network Volume. Size it for your purchase using the table below.
| Your purchase | Minimum volume size |
|---|---|
| Image Pack | 30 GB |
| Video Pack | 60 GB |
| Image + Video Pack | 90 GB |
Set up your RunPod credentials
In the RunPod console, open Credentials and create two secrets with these names exactly.
| Secret name | Value |
|---|---|
| _huggingface_token | Your Hugging Face token: huggingface.co, Settings, Access Tokens, type Read |
| _civitai_token | Your CivitAI API key: civitai.com, Account settings, API Keys |
Switch to the Registry credentials tab and click Add.
| Field | Value |
|---|---|
| Name | Ahegao Registry |
| Username | x |
| Password | Your license key from ahegao.red, without extra spaces |
Deploy the pod from the official template
Each pack ships a link that opens the RunPod deployment page with the right GPU preselected. Use the link from the setup PDF, included with your purchase.
- Click Select registry authentication and choose the Ahegao Registry credential you created.
- Click Select Network Volume and choose the volume from step 1.
- Click Deploy.
First start and model provisioning
The pod starts and downloads all required models onto the volume. This happens once, and interrupted downloads resume on restart.
Wait until the log reports "To see the GUI go to: http://0.0.0.0:8188" before you connect.
Connect to ComfyUI
Use the Connect button, then HTTP Service on port 8188. Your purchased workflows appear under the Workflows panel. Open one, replace the example prompt, and press Run.
If ComfyUI does not load, add a trailing slash to the proxy URL and reload.
Managing costs
Stop the pod when you finish generating. A stopped pod costs no GPU time and keeps everything in place.
Models and outputs stay on your network volume. Stopped pods cost storage only.
Updates
Updates ship weekly. Deploy a new pod from the official template, or pull the latest image for your pack. The network volume keeps all models and outputs.
Troubleshooting
| Symptom | Resolution |
|---|---|
| RunPod rejects the image pull as unauthorized | Check the registry credential: username x, password is the license key without extra spaces |
| Template link does not open | Search the RunPod Hub for ahegao_red_image, ahegao_red_video, or ahegao_red_image_video |
| Provisioning stops with DNS errors | Stop and start the pod. RunPod reschedules it onto different host hardware |
| Disk-full warnings | Your network volume is smaller than the pack requires. Resize it and restart |
| Output quality looks distorted | The shipped workflows are pre-calibrated. Do not modify sampler, CFG, or LoRA settings |
Support
Write to support@ahegao.red if the steps above do not solve your problem. Include your license key, the pack you bought, and the pod logs.
The printable copy of this setup ships as instructions-runpod.pdf on ahegao.red.
Adults only (18+). All sample output is synthetic, computer-generated fiction. Last verified October 2026.