Documentation · Development build 0.1.0

Project. Paint.
Make it yours.

A visual reference for AI projection and texture painting. Follow the actual panels from installing a backend and connecting, through cameras, ControlNet, style reference, masks, lighting, material tools, upscaling and export — plus image-to-3D generation.

Buy ArtBrush ↗Start the guide ↓Install & models 3D Generation Guide

01 / PANEL GUIDE

Your first textured model

Start with a UV-mapped model and one projection camera. Establish one useful view, then add coverage, repair the texture and export.

ArtBrush workspace with projection cameras, viewport and AI texture generation panel — The ArtBrush workspace — projection cameras on the left, the viewport in the center, AI texture generation on the right. The menu bar adds Reload, Import 3D, Generate 3D Model and Install; the toolbar starts WebUI / ComfyUI and opens this documentation.Open full size ↗

Open Install once to set up a backend and download the models you need (see 02).
Press Run ComfyUI (or Run WebUI) in the toolbar, then Connect in the AI Server group.
Choose Import 3D (or File → Import 3D Model…), frame the model with F and use Align to View on a camera.
Pick a Quick Setup scenario, describe the surface and press Generate & Apply. Inspect the result from several angles.
Add views for uncovered areas (or use Total Object Multi-Angle), blend the cameras, then repair seams with a mask.
Generate PBR / ORM maps if needed, save the project and export textures or the textured model.

Development guide · Based on the current ArtBrush 0.1.0 source. Interface labels and available workflows may change during development.

02 / PANEL GUIDE

Install, launch & connect

ArtBrush is a standalone desktop workspace. AI generation runs on a local ComfyUI or WebUI Forge Neo backend that ArtBrush can install and start for you.

Install — Backends & Models

One dialog installs a backend and downloads the model files it needs.

1. Backends: Install WebUI… / Install ComfyUI run the bundled installer from artbrush/server/ in a console. The button turns green when that backend is installed.
2. Model folder: One shared folder used by both backends. Browse to an existing ComfyUI / WebUI models folder to reuse it, or use Default.
Civitai API key: Only needed for models that require a Civitai login.
3. Models: Grouped by Checkpoints, Text Encoders, VAE, ControlNet, Style Reference, Skin LoRA, Weathering & Decal LoRA and PBR / 3D / Anti-Shade. Each row shows the backend, a tier (Core / Recommended / Optional), the target folder and Installed / Missing status.
Selection & download: Core, Recommended, All, None and WebUI-only / ComfyUI-only quick selections; Download Selected fetches the checked files, Cancel stops it and Refresh rescans the folder.

ArtBrush panel: Run backends from the toolbar — Interface reference · Open full size ↗

Start or stop a backend without leaving ArtBrush.

Run WebUI / Run ComfyUI: Starts WebUI Forge Neo or MotionSmith ComfyUI in the background. While running, the button reads Stop ….
Console (C): The small button beside each backend shows or hides its console window — useful for reading backend errors.
Documentation: Opens this guide.

AI server connection

Connect the backend before selecting its installed models.

Backend: ComfyUI (WebSocket + API Graph) or WebUI / SD-Forge (REST API). They expose different workflow capabilities.
Host / Port: ComfyUI defaults to 127.0.0.1:8188; WebUI / Forge to 7860.
Connect / status: The green Online message confirms the endpoint was reached; it does not confirm every required model or node is installed.

Reload: The Reload menu restarts ArtBrush through its launcher — handy after installing models or nodes.
Development launch: On Windows, run.bat uses uv run --locked python -m artbrush, or the existing .venv. The source checkout requires Python 3.12 or newer.

03 / PANEL GUIDE

Import & prepare the model

Use Import 3D or File → Import 3D Model… to load OBJ, GLB, glTF or FBX.

Untextured character imported into ArtBrush — Before — an imported, untextured character with scene lighting.Open full size ↗

Character textured with Total Object generation in ArtBrush — After — the same model textured with one Total Object generation (SDXL · Depth).Open full size ↗

UV layout: Use a model with a suitable UV layout. Inspect its coverage and seams before building a multi-view texture.
FBX: FBX import/export uses an external backend. If ArtBrush reports it missing, use OBJ or GLB.
Projects: New Project, Open Project (.artbrush)… and Import SPZ Project (.spz, partial)… are in the File menu.
View menu: Frame Model (F), Align Camera to View, Toggle Projection Preview, Floor Grid and Reset Camera.

04 / PANEL GUIDE

Cameras, gallery & projection

Organize generated views, select camera poses and blend projections before touching the brush tools.

Generated Views & imported maps

Every generation is kept as a card tied to its camera.

Result cards: The title identifies camera and iteration. Active on Model marks the applied result; Saved View is a stored alternative.
Apply / View / Hide: Apply restores that record to the model; View restores its camera pose; Hide / Show toggles only that view's projection layer on the model.
Download / Delete: Save the card's image or remove the record.
UV Texture… / Normal Map… / Projection…: Import UV-space color or normal maps (select all files of a UDIM set together), or an image for the active camera.
Background ▾: From File, From View, Generate or None — see Background & Atmosphere.
Bake All Textures: Flattens the current model texture into one BaseColor record and removes all generated images (camera images, AO, UV maps) and their undo steps to free memory and VRAM. Background and normal maps stay. Cannot be undone.
Clear All: After confirmation removes gallery images and their texture/camera contributions, normal maps and background. Ctrl+Z restores it; the mesh stays.

Projection cameras

The camera list is the starting point for a multi-view texture.

Entries: Select a row to edit it. [Has Image] / [Masked] describe attached inputs; w is the blend weight.
Add (+) / Remove (-): Create a camera or remove the selected one.
Cardinal 4-View (Preset): Replaces the camera set with Front, Right, Back and Left after confirmation.

Selected Camera Properties

Controls for the selected projection camera.

Projection Enabled / Mirror: Enable its contribution; Mirror (X Symmetry) also projects onto the mirrored side.
Weight / Opacity: Relative weight in multi-camera blending, and strength of its projected image.
Align to View / Look Through: Copy the viewport pose to the camera, or move the viewport to the camera.
Source Image / Contribution Mask: Load or clear the camera's image and its contribution mask.

Projection Settings

Refine the latest projection layer without generating again.

FAR: Angle falloff: how quickly influence fades as the surface turns away from the camera.
VIS / SFT: Visibility of the latest projection over the texture below, and edge softness of the camera-image border.
HUE / SAT / VAL / CST: Hue, saturation, brightness and contrast of the projected image.
Reset to Default: Restores the settings; changes reapply the stored projection without a new AI request.

Multi-Camera Blending & Baking

Combine camera projections into the model's UV texture.

Gamma: Angular weighting: surfaces facing a camera get more influence than grazing angles.
Edge Margin: Image-edge margin used by projection blending.
Blend All Cameras: Bakes all enabled cameras with their weights, images and masks.
Bake Active Camera Only: Bakes the selected camera alone to inspect one contribution.

AO (Ambient Occlusion) & edge padding

Bake local AO on a UV-mapped mesh and pad texture borders.

Sample Count / Distance / Strength: Number of sampling directions, ray distance (0 = automatic from model size) and darkening amount.
Empty Areas Only: Fills only unpainted/transparent texels with AO.
Bake & Apply AO: Computes and applies AO to the texture tiles.
8 px / 16 px: Edge dilation: extends painted edges into empty UV space (also in the Bake menu).

Left dock tabs

The left dock switches between three panels.

Projection Cameras: Gallery, cameras, projection, blending and AO (this section).
Brush & Mask: Painting, masks and history — see Brushes & masks.
Scene Lighting: Lights, AI lighting guidance and background — see Lighting.
Docking: Panels can be floated or closed with the buttons in their title bars.

05 / PANEL GUIDE

Models, Quick Setup & LoRA

Model files belong to your AI backend; the Install dialog downloads them into the shared model folder. Quick Setup then selects installed models and applies scenario settings.

Quick Setup

Pick a model family, then a scenario from its menu.

Families: Flux 1 (Dev / Schnell), SDXL, SD 1.5, Qwen-Image, Z-Image and Tools. Clicking a button applies its first scenario; the arrow opens the full list.
Scenarios: SDXL: Depth, Canny, Depth + Canny, No ControlNet. SD 1.5 adds Normal Map and Depth + Normal. Z-Image: Depth, Depth + Canny, No ControlNet. Qwen-Image: Depth. Flux 1: Dev or Schnell with depth guidance.
Tools → Remove Shadow: Selects the Shadow_R / Anti-Shade workflow; press Generate after selection.
Backend limits: WebUI ControlNet works with SD 1.5 / SDXL only; other scenarios are disabled there and marked ComfyUI only.
Download models…: Appears when a scenario finds no model and opens the model guide.

ComfyUI Model & FreeU

The model controls the selected scenario filled in.

Model Family / Refresh Models: Select the loading path (checkpoint or diffusion model + encoders), then refresh lists after installing files.
Checkpoint / VAE: Installed base model; Checkpoint VAE uses the VAE inside the checkpoint.
Depth ControlNet: A depth model compatible with the family, e.g. controlnet-union-sdxl promax.
Sampler / Scheduler: Sampling algorithm and noise schedule (dpmpp_2m / karras here).
FreeU (B1 B2 S1 S2): Sharper texture detail and fewer muddy flat areas on SD 1.5 / SDXL; enabling it loads recommended values for the checkpoint.

LoRA Selection

Apply up to two LoRAs to every generation.

LoRA 1 / LoRA 2: Choose an installed LoRA or None (Disabled). Works with ComfyUI and WebUI / Forge.
Model / CLIP Strength: Influence on the diffusion model and on text conditioning.
Compatibility: Match the LoRA to the base family (SDXL LoRA with an SDXL checkpoint, and so on).

ComfyUI Workflow

Use the built-in graph or load your own API graph.

Load Workflow: Loads ComfyUI API-format JSON; a normal UI-layout workflow is not the same format.
Reset Template: Returns to the built-in graph for the current workflow mode.
Status: Built-in graph means the ComfyUI Model controls drive generation. A Quick Setup recipe also returns to the built-in graph.

Model families: FLUX.1 dev / schnell (diffusion model + T5XXL + CLIP-L + ae VAE), SDXL and SD 1.5 checkpoints, Qwen-Image and Z-Image-Turbo (diffusion model + their text encoder and VAE). Keep encoders, VAE and ControlNet in the same family.
Folder reference: Under the model folder: checkpoints/, diffusion_models/, text_encoders/, vae/, controlnet/, loras/, ipadapter/, clip_vision/, style_models/.
WebUI / Forge models: WebUI uses the same shared folder when installed through ArtBrush; select its checkpoint and modules in the model controls after connecting.

Model sources: SDXL ↗ · FLUX.1 dev ↗ · ComfyUI model paths ↗. Follow each publisher's terms.

06 / PANEL GUIDE

Texture workflow modes

Choose the target operation before selecting prompts and generation strength.

Read the selected workflow

The description under Workflow Mode explains the current edit target.

Projection Masking: Edits only the active camera's contribution mask. Other camera masks and the overall texture are preserved.
Status messages: Read the mode description together with any prerequisite message before generating.

ArtBrush panel: Seven texture workflows — Interface reference · Open full size ↗

Each mode prepares a different input or target region.

Projection Masking: Controls the selected camera's projection contribution.
Total Object: Generates the whole model as seen from the current view. On a textured model, Re-do regenerates from its render; Blur / Edge soften the view's region into the existing texture.
Total Object Multi-Angle: Generates the selected cameras (Front, Right, Back, Left) in sequence and blends them in one pass.
Inpaint with Color Guidance: Sends painted color as the init image and the edited area as the mask; the result is mapped only there.
Soft Inpainting / Seam Repair: Keeps the baked texture as init and sends an uncolored mask — for seams and defects.
Fill Transparent / Empty Areas Only: Fills only alpha = 0 texels; black paint counts as filled.
AI Delighting & Shadow Removal: Cleans baked lighting, harsh shadows and AO with an AI delighting workflow.

re-do / blur / edge dials: Re-do sets regeneration denoise on supported paths; Blur feathers inward from the target region; Edge softens its boundary. Drag, use the wheel or double-click to type.

07 / PANEL GUIDE

Prompts, sampling, size & upscale

Values shown in the captures belong to one SDXL Lightning example; Quick Setup replaces them per family.

Generation

Describe the texture and control sampling.

Positive / Negative Prompt: What to create and what to avoid (when the family uses negative conditioning; FLUX reads the positive only).
Front / Back / Top / Bottom: Add view words such as front view, centered, straight-on angle to the prompt.
+ Add Negative: Appends the built-in negative template.
Steps / Seed / CFG: Sampling iterations, seed and conditioning strength; Random requests a new seed each generation.
Prompt presets / Styles: Categorized prompt presets (e.g. Bionic Ceramic Android) fill the prompt; Styles applies a style from styles.csv.

Apply

Run generation once scene and backend inputs are ready.

Matte albedo (no reflections): Adds matte/diffuse terms to the prompt and specular highlights, glossy reflections and glare to the negative, so light reflections are not painted into the texture.
Generate & Apply: Runs the selected workflow and bakes the result to the model.
Progress / message: Shows progress and the result, e.g. “AI texture generated and baked to model!”. Still inspect the surface.

Resolution / Ratio & Texture Size

Generation image size is separate from the UV texture size.

Ratio Preset / Width / Height: Square, portrait, landscape and cinematic presets, or custom 64–4096 px in 64 px steps.
Texture Size: UV texture resolution used for baking: 1024, 2048 or 4096 (default). Close-ups need more texels; existing smaller textures are upscaled.

Upscale · Hires Fix

A second sampling pass on the upscaled image.

Method: Hires Fix or Ultimate SD Upscale.
Upscaler: An installed upscale model (e.g. 4x-ClearRealityV1) or a plain resize.
Multiplier / Denoise / 2nd Pass Steps: Scale factor, how much the second pass may change the image, and its steps.

Upscale · Ultimate SD Upscale

Upscales, then redraws the image tile by tile — ComfyUI only (skipped on WebUI).

Upscaler / Upscale By / Passes: Upscale model or Lanczos / Nearest resize, factor, and 1 or 2 passes (2 × 2x gives 4x).
Steps / CFG / Denoise / Sampler / Scheduler: Tile redraw settings; Same / Same as generation reuse the main values.
Mode / Tile Width / Tile Height / Mask Blur / Tile Padding / Batch: Tiling pattern and tile size, blending and how many tiles run at once.
Seam Fix / Seam Denoise / Width / Mask Blur / Padding: Optional second pass along tile seams.
Force Uniform Tiles / Tiled Decode: Keep tiles the same size; decode the VAE in tiles to save VRAM. The footer shows the final size.

08 / PANEL GUIDE

ControlNet & Style Reference

Build guidance from camera data, and optionally steer the look with a reference image.

Depth Map Settings & ControlNet 0

Prepare the depth input, then configure the first slot. ControlNet 1, ControlNet 2 and Style Reference are collapsible below it.

Bright / Contrast / Blur / Edge: Adjust the depth map (50 = unchanged); edge-preserving blur and the threshold for keeping sharp depth boundaries.
Blur Edge / Inside: Uniform blur that also softens boundaries; Inside keeps it on the model so the background stays black.
P / B / C / LOW: Prompt priority, Balanced, ControlNet priority and low-memory mode (WebUI options).
Sources: Vertex colors, View (rendered texture and lighting), Depth, Normal, No Image (WebUI inpaint) or Custom File. Click the preview to refresh.
Preprocess / Model: None passes the prepared map; ComfyUI supports None and canny. Automatic picks a compatible installed model.
Start / End / Weight / res: Guidance interval, strength and preprocessor resolution (×½, ×1, ×2).

ControlNet 1 · Normal input

A second, independent slot — here guiding surface orientation.

Normal: The camera's colored surface-orientation map, not a UV normal texture.
Independent settings: Own enable state, priority, preprocessor, model, start/end, weight and resolution.
Example: Automatic model, Weight 0.80. Match the model to the base family.

Style Reference (IP-Adapter / Redux)

Transfer the look of a reference image to the generation.

Choose Reference Image…: Loads the image whose style, colors and materials guide the result.
Redux / IP-Adapter: FLUX.1 uses Redux (flux1-redux-dev in style_models/); SD 1.5 / SDXL use IP-Adapter (ipadapter/). On WebUI it runs as a ControlNet IP-Adapter unit.
CLIP Vision: sigclip for Redux; CLIP-ViT-H-14 for IP-Adapter.
Weight / Start / End: Strength and the denoising interval where the reference applies.

09 / PANEL GUIDE

Brushes, masks & history

Direct color, the soft inpaint mask and color guidance are separate editing layers.

Tools

Choose the layer you intend to edit before painting.

3D Navigation (Orbit): Viewport navigation instead of painting.
Texture Paint: Brush (Color) / Erase Paint edit the texture directly.
Inpaint No-Color (Soft Mask): Paint Mask / Erase Mask edit the mask used by Soft Inpainting.
Inpaint Color (Color Guidance): Color Paint / Erase Color Paint edit the guidance layer for color-guided inpainting.
UV Island Fill: Fills a whole UV island with the current color.

Brush Settings

Tune the stroke independently of AI sampling.

Size (Radius) / Hardness: Brush footprint and falloff.
Opacity (1–100%): Application strength.
Color / Eyedropper: Pick a color or sample the model (Alt + click).
Tablet: Pressure → Size and Pressure → Opacity.

Mask Operations

Edit the whole mask at once.

Invert Mask: Swaps selected and unselected regions.
Clear Mask: Removes the painted mask.

History (Undo / Redo)

History reports the most recent supported edit.

Undo / Redo: Ctrl+Z; Ctrl+Shift+Z or Ctrl+Y. A disabled button means nothing is available.
Last Action: Label of the latest step, e.g. AI texture generation.

Painting shortcuts: Shift + left drag paints in the inpaint tools; left drag alone orbits. Alt + click samples a color. X toggles a paint/erase pair.
Local repair: Choose the inpaint mode, paint its region or color guidance, set re-do / blur / edge and Generate. Check the join from more than one angle.

10 / PANEL GUIDE

Lighting, guidance & background

Separate the surface color from the way you inspect and guide it.

Lighting Setup & Presets

Real-time lighting for display and optional AI guidance.

Enable Scene Lighting: Toggles the scene-lighting system.
Preset: Named setups; Custom Setup… means the controls were edited manually.

Lighting preset menu

Presets fill the lighting controls; they never change the mesh or texture.

Studio / Dramatic / Sunset: Studio (Soft), Dramatic Rim and Sunset Warm.
Cool / overhead: Cyberpunk Cool and Overhead High-Key.
Neutral Flat (Delighted): A flat display preset for judging color; it does not run AI shadow removal.

Key Light (Main Direction)

Set the main light relative to the model.

Azimuth / Elevation: Horizontal angle around the model and vertical height (45° / 35° here).
Intensity / Light Color: Key-light strength and RGB color.

Ambient Fill & Rim Light

Tune the fill separately from the key light.

Ambient Fill / Tint: Base illumination and its color.
Rim Highlight: Edge highlight strength — a display/guidance effect, not an emissive map.

AI Generation Guidance

Use scene lighting as an input cue while keeping it out of the stored color.

Apply Lighting to Generation (Light Isolation): Sends a lit view as input, then divides the captured shading out of the returned color.
Light-guided Denoise: Img2img strength with light isolation: lower keeps the lit input closer.
Auto-inject Lighting Prompts: Adds lighting words computed from the active camera; the live guidance box shows them (e.g. “three-quarter key light from right”).

Background & Atmosphere

A backdrop gives context without becoming the UV texture.

Viewport Atmosphere / Brightness: Viewport backdrop tone and its brightness percentage.
None / From File… / From View / Generate…: Clear, load, reuse the active camera's last image, or generate one from the text field.
Send to AI with the camera view: Includes the backdrop in View / img2img inputs; depth, normal and masks keep a black background.
Reset to Default (Studio): Resets scene-lighting values.

11 / PANEL GUIDE

Skin, decals, weathering & PBR

Optional collapsible groups; each needs its matching model family and backend files.

LensArt Ultra Skin Detailer

Skin-detail LoRA plus matte-skin prompt edits.

Preset / LoRA / Download: Choose a preset (FLUX, SDXL or SD 1.5) and its LoRA; Download opens the model page.
Model / CLIP Strength: LoRA influence.
Apply Matte Skin Prompt: Adds positive/negative phrases and removes glossy-skin words from the positive prompt.
Override CFG / Guidance: Only applies when its checkbox is on.

Decal & Stencil

Load or generate a decal and project it from the current view.

Decal LoRA / Checkpoint: The checkpoint is selected automatically from installed models for the LoRA family; Installed shows the file is present.
Category / Presets / Decal Prompt: Prompt presets or custom text.
LoRA Strength / Opacity: Generation influence and how strongly Apply Decal places it.
Remove Background: Makes the plain/gradient background transparent; the number is the color-distance tolerance.
Size / Position X Y: Decal height as a fraction of the view, and its offset.
Open / Generate / Apply Decal: Load an image, generate one, then project it as an adjustable layer.

Weathering & Wear Slider

Bias generation with a weathering LoRA.

Weathering Preset / LoRA: SD 1.5, SDXL or FLUX presets (aging, cracks, rust…) and the installed file.
Wear / Aging: LoRA strength — not a deterministic damage percentage.
Use Prompt Suffix: Appends the prompt / negative suffix fields; Category and Presets fill them.

ORM-PBR Map Generation & Export

Estimate material maps from the texture and export them.

CHORD Model / Generate Maps: Needs ComfyUI with CHORD nodes (Ubisoft La Forge). Generates normal, roughness, metallic and optional height from the active texture.
Generate Height / Invert Normal Y (Gen): Height derivation and green-channel flip during generation.
Bake AO: Bakes AO and stores it for export and ORM.
PBR Maps / ORM Map / UDIM-BaseColor / AO Map: Choose what Export Selected Maps writes. ORM packs R = AO, G = Roughness, B = Metallic.
Normal: DirectX (Y-): Flips green on export (Unreal / DirectX). Avoid flipping twice.
Export Selected / All Project Maps / Camera Views: Checked maps only; everything (renders, UDIMs, PBR, ORM); or the active camera's RGB, normal, depth, mask, edges and raw depth.

12 / PANEL GUIDE

Save projects & export models

Save an editable project first, then export maps (see ORM-PBR) or the textured model.

Editable projects: Save Project (Ctrl+S) / Save As… write a .artbrush manifest with a companion .artbrush_data folder. Keep both together.
PNG & UDIM: Export Baked Texture (PNG)… and Export UDIM Tiles (PNG)… in the File menu.
Textured models: Export Model as GLB, OBJ + MTL + PNG or FBX (FBX needs its external backend). Check material assignment and normal convention in the destination.
AI input maps: AI → Export AI Input Maps… saves the active camera's ControlNet inputs.
SPZ: Import SPZ Project (.spz, partial) is a separate, partial importer — not a texture export.

13 / PANEL GUIDE

Generate 3D Model

Create a mesh from an image inside ArtBrush, then send it straight to texture painting.

Image-to-3D workspace

The Generate 3D Model menu opens a dedicated workspace; Back to Texture returns.

Workflows: Pixal3D, Pixal3D MultiView, TRELLIS.2, TRELLIS.2 LowPoly and TRELLIS.2 RefineMesh (ComfyUI).
ComfyUI Models: Diffusion model, Shape VAE, DINOv3 and background removal (BiRefNet) for the chosen workflow.
Workflow Settings: Preset (e.g. Medium), seed randomization and per-node settings: samplers, resolution, CFG, decimation, remesh, UV unwrap and more.
Generate / Send to Texture Painting: Generate 3D Model builds the mesh; Send to Texture Painting loads it into the projection workspace.

Full image-to-3D guide
Workflow choices, model requirements, mesh examples and GLB transfer.

3D Generation Guide

14 / PANEL GUIDE

Troubleshooting & shortcuts

Resolve connection and input problems before changing generation parameters.

Cannot connect: Start the backend (Run ComfyUI / Run WebUI) and open its console (C) to read errors. Check the backend type, host and port.
Empty model selectors: Open Install, check Missing rows and Download Selected, then Refresh Models. For FLUX also check T5XXL, CLIP-L and the VAE.
Scenario disabled: WebUI supports SD 1.5 / SDXL ControlNet only. Connect ComfyUI for Flux, Qwen-Image, Z-Image, Ultimate SD Upscale and CHORD.
Generation is blocked: Read the status message: mask workflows need a mask, Multi-Angle needs cameras, Style Reference needs an image and matching models.
Seams or missing coverage: Use Look Through on each camera, check enable state and mask, blend again; add views or use Seam Repair.
Memory / VRAM: Bake All Textures drops stored generations; Tiled Decode and LOW reduce VRAM; lower Texture Size if needed.
Keyboard reference: F frame · Ctrl+I open projection image · Ctrl+S save · Ctrl+E export baked texture · Ctrl+Z undo · Ctrl+Shift+Z / Ctrl+Y redo · X brush/eraser pair · Alt+click eyedropper.

Need help? Visit MotionSmith Support ↗. Include your ArtBrush build, backend, selected models and the exact status message.

MotionSmith ArtBrush Documentation