Documentation
Everything you need to get started with Noisegate.
Overview All DCCs
Noisegate is a rendering bridge that allows artists to run AI workflows directly inside professional 3D applications. The system connects Houdini, Blender, and Cinema 4D with ComfyUI through a unified launcher and renderer integration.
Instead of switching between multiple tools, Noisegate enables artists to run AI workflows directly inside their existing production environment.
System Components
- Noisegate Launcher — The central hub that manages the ComfyUI engine, workflows, updates, and account access.
- Noisegate Render Plugins — Native integrations for each DCC application (Houdini, Blender, Cinema 4D).
All DCC plugins communicate with ComfyUI through the launcher.
Noisegate Launcher All DCCs
The Noisegate Launcher is the central hub of the system. It allows users to:
- Install and run ComfyUI
- Manage workflows
- Connect 3D applications
- Update the software
- Manage accounts and subscriptions
The launcher is structured into several tabs: Engine, Setup, DCC Applications, Workflows, Updates, and Account.
Engine Tab All DCCs
The Engine tab is the first page displayed when the launcher starts.
Features
- Start / Stop ComfyUI
- Check engine connection status
- Change the port (optional, default:
8188) - View engine logs
The log panel shows the full startup process of ComfyUI, including model loading and node registration.
Setup & Installation All DCCs
The Setup page allows users to install ComfyUI.
- Click the Install button to start the installation
- A log window displays the progress in real-time
- Installation typically takes 10 to 15 minutes
- The installation path is validated automatically
Users can manually verify the installation using Validate Path if needed.
DCC Applications All DCCs
This section allows users to register their installed 3D applications.
Supported Applications
- Houdini — SideFX Houdini 19.5+
- Blender — Blender 4.0+
- Cinema 4D — Cinema 4D 2024+
Users must provide the application paths. Each path can be validated to confirm that the application is detected correctly.
Workflows All DCCs
The Workflows tab provides access to available AI workflows.
Features
- Refresh the workflow list
- Download workflows
- Install missing custom nodes automatically
Workflows are centrally maintained and updated with the latest models and ComfyUI API nodes.
Coming Soon
- Private workflow uploads
- Team workflow sharing
- Workflow hosting
Teams will be able to share workflows through the launcher while keeping them private within the group.
Updates All DCCs
Displays the current software version. Users can check for updates and install them directly from the launcher.
Noisegate auto-updates on launch — always stay on the latest version.
Account All DCCs
Displays user account information:
- Subscription tier
- Expiration date
- Partial license key reference
Houdini — Setup Houdini
Noisegate integrates with Houdini through a custom ROP (Render Output) node.
- Create a ROP Network (or navigate to
/out) - Open the Tab menu inside the network
- Select Noisegate Renderer
The Noisegate ROP appears as a standard render node and can be connected to other ROPs in the dependency graph.
Houdini — Workflow & Settings Houdini
Loading a Workflow
- Select a workflow from the dropdown
- Click Load Workflow
After loading, all workflow parameters are populated in the node interface.
Settings Panel
- Prompt — The text prompt sent to the AI model
- Seed — Controls randomness (same seed = same result)
- Prompt Refinement — Optional prompt enhancement
- System Prompt — Instructs the AI model's behavior (editable per workflow)
Houdini — Rendering Houdini
Render Configuration
- Camera — Select the scene camera to render from
- Resolution — Aspect ratio presets or custom resolution
Output Settings
- Output Directory — Where rendered images are saved
- Filename — Naming pattern (supports
$F4frame padding)
Click Render on the ROP node to start. The render process captures the camera view, uploads it to ComfyUI, runs the workflow, and downloads the result.
Houdini — Image Inputs Houdini
Image inputs depend on the selected workflow and may include:
- Camera Render — Standard beauty pass from the scene camera
- Depth Pass — Normalized Z-depth buffer
- Normal Pass — World-space surface normals
- None — Skip this input
- File — Use an image file from disk
Each LoadImage node in the workflow appears as a configurable input slot. The pass type can be set independently per input.
Houdini — Output Passes Houdini
Workflows may define multiple output passes. Each SaveImage or PreviewImage node in the workflow appears as a toggleable output pass.
Enable or disable individual passes to control which results are downloaded.
Houdini — Render View Houdini
Noisegate adds a custom render viewer panel to Houdini.
Creating the Render View
- Open the Pane Tab menu (plus icon)
- Select New Pane Tab Type
- Choose Noisegate Render View
Features
- Camera selection
- Render preview with zoom and pan
- Image scaling
- Render history browser
- Video generation workflow support
Houdini — Video Generation Houdini
Noisegate supports AI video generation workflows (e.g., Kling, WAN, LTX Video).
Workflow
- Render an image using an image workflow
- Select a video workflow from the dropdown
- Enable Lock Style to use the rendered image as style reference
- Define the frame range for animation export
- Animate the camera in the scene
- Click Render
The system uses the rendered image as a style reference and the camera animation as motion input. The exported MP4 is uploaded to ComfyUI and processed by the video model.
Blender — Setup Blender
Noisegate appears as a render engine in Blender.
- Open Render Properties in the Properties panel
- Change the Render Engine dropdown to Noisegate
If the Noisegate Launcher is running and ComfyUI is connected, the interface displays Connected with the server URL.
Blender — Workflow & Settings Blender
Loading a Workflow
- Select a workflow from the dropdown
- Click the Refresh icon to load it
After loading, the following panels appear in Render Properties:
- Image Inputs — Configure camera passes or file inputs for each LoadImage node
- Output Passes — Toggle which output nodes to download
- Parameters — All tweakable workflow parameters (prompt, seed, model, system prompt, etc.)
- Output — Output directory, filename, and animation frame range
- Render — Start rendering
Resolution
If the workflow contains resolution nodes (e.g., EmptyLatentImage, Flux API nodes), a resolution preset dropdown appears with common aspect ratios. Custom resolutions can also be set manually.
Blender — Rendering Blender
Click Render in the Render panel to start. The process:
- Renders the active camera (beauty, depth, or normal pass depending on input configuration)
- Uploads the rendered image to ComfyUI
- Applies all parameter overrides (prompt, seed, resolution, etc.)
- Submits the workflow and polls for completion
- Downloads the result to the output directory
- Displays the result in Blender's Image Viewer
Progress is shown in the Render panel. Press Esc to cancel a running render.
Render History
All rendered results are tracked in the History panel. Select an entry and click Load to view it again. Video results support timeline scrubbing.
Cinema 4D — Setup Cinema 4D
Noisegate appears as a standard renderer in Cinema 4D.
- Open Render Settings (
Ctrl+B/Cmd+B) - Change the Renderer dropdown to Noisegate
The Noisegate interface appears in the Render Settings panel.
Cinema 4D — Interface Cinema 4D
Sections
- Prompts — Workflow selection, system prompt, and prompt input
- Output — Export directory and filename prefix
- History — Render history storage
All workflow parameters are exposed as native Cinema 4D interface elements.
Cinema 4D — Rendering Cinema 4D
Rendering works like any native Cinema 4D renderer:
- Shift + R — Render to Picture Viewer
- Render icon in the toolbar
- Picture Viewer — Results appear directly in Cinema 4D's Picture Viewer
The render process follows the same pipeline as other DCCs: camera capture, upload, workflow execution, and result download.