CNC Carving

Photo-based 3D relief, or precise V-bit text carving for signs and plaques

Working…
Why?Pick the ZIP a previous CNC Relief job download gave you. Every setting on this page (wood, wood width, carving depth, relief sliders, borders, backing, masks, laser overlay, and Direct G-code toolpath settings) gets refilled from that job's _settings.json so you can change just one or two values and re-generate. The original photo also gets pulled back out of the ZIP and re-selected automatically when possible — re-select it manually below if not.
CAM Software
The STL output works with any CNC machine. This selects which step-by-step instructions are in your PDF guide.
Material & Dimensions
💡 Enter your actual measured thickness or select a common size from the dropdown.
Selecting specific wood may help with material-specific optimizations.
Fixed = specific depth values. Percentage = depth scales with material thickness.
This is your physical wood size. Enter the actual measured width (use the dropdown for standard lumber). ℹ️ Note: Carving area will be slightly smaller than wood size (0.5" border on all sides for edge safety). Height is calculated from the photo's aspect ratio after upload.
Photo

Select the type of image you're uploading

Color Photo: AI will analyze your image to understand 3D structure (noses protrude, backgrounds recede, etc).

Choose which depth estimation model to use

V3: DA3MONO-LARGE monocular specialist. V2: DA2-Large via HuggingFace. Both Apache 2.0 (commercially safe).

🖼
Drop photo here or click to browse
ℹ️ Supported: JPG, PNG, TIFF (including 16-bit greyscale), BMP, WebP.
AI Depth & Relief Tuning

Fine-tune the AI depth estimate and edit height masks in the full-screen editor.

Laser Engraving (optional)

Generate laser G-code to char edges on your carved relief. G-code goes to your CNC controller (not separate laser software).

Creates .nc file for your CNC controller. Laser follows 3D surface to maintain focus.

Flat laser G-code or a flat UV-print PDF for the back of a carving or standalone piece. Text, handwriting images, maker's marks, dedications.

Upload a photo to see the depth map preview