State-of-the-art image and video enhancement using NVIDIA RTX Video SDK. It executes up to three sequential passes (Denoise, Deblur, and Upscale) in a single node, processing frame-by-frame to keep VRAM usage predictable and low.

Refine: Independent Denoise and Deblur passes.
Upscale: AI-powered VSR and High Bitrate upscaling.
Smart Sizing: Multiple resize modes including Constant Megapixel targets.
Efficiency: Frame-by-frame processing for minimal VRAM usage.

Full documentation →

📐 Resolution Scale Calculator

The DaSiWa Scale Calculator provides mathematically precise resolution management for high-performance video models. It uses a Constant-Area Square-Root method to ensure that your GPU VRAM usage remains stable regardless of the aspect ratio.

Full documentation →

🎛️ Node Status Switch

The DaSiWa Node Status Switch lets you mute or bypass any node in your workflow using a single toggle. Targets are registered by wiring their outputs into the switch's input slots, which grow dynamically as you connect more nodes (up to 99).

Full documentation →

Quick start:

Add a DaSiWa Node Status Switch to your workflow
Drag any output from the node(s) you want to control into the switch's target_01 input — new slots appear as you connect more
Set action to mute or bypass and configure trigger_on to taste
Toggle enabled directly on the switch

💾 Metadata Image Saver (Civitai Ready)

The DaSiWa Metadata Image Saver ensures your images are fully compatible with Civitai, Hugging Face, and other galleries by embedding A1111-style metadata. It automatically detects LoRAs used in the workflow and supports dynamic filenames.

Civitai Compatibility: Writes the standard parameters block for auto-parsing of prompts and resources.
LoRA Detection: Scans your workflow and appends <lora:name:weight> triggers automatically.
WebP Support: Full "Drag-and-Drop" workflow reconstruction support for both PNG and WebP formats.
Dynamic Filenames: Use placeholders like %seed%, %date%, %model%, %width%, and %height%.
Privacy: Toggle workflow JSON embedding to share images without exposing your full graph.

Full documentation →

🛠️ Installation

Manual install

Activate your venv inside your ComfyUI folder

Clone this repo into your custom_nodes folder:

git clone https://github.com/darksidewalker/ComfyUI-DaSiWa-Nodes

Install all dependencies:
```
pip install -r requirements.txt
```
Requirement: NVIDIA RTX GPU with drivers 530+. (Windows users may need the NVIDIA Broadcast SDK; Linux usually works out-of-the-box with the pip package).
Restart ComfyUI.

Use ComfyUI-Manager

Search for DaSiWa-Nodes and install.

Credits

The RTX implementation in this collection is based on the excellent work by Deno2026/comfyui-deno-custom-nodes.