ComfyUI

WORKFLOW SEARCH ENGINES: ComfyUI - GitHub - GitHub2 - RunComfyUI for Nodes - ComfyUI/Reasoning - 

Search for ComfyUi Height Map Demo Site - STL conversion tools

Depth Anything - https://github.com/PozzettiAndrea/ComfyUI-DepthAnythingV3 

 Best Workflow: ComfyUI-DepthAnythingV3 + Depth2Mesh

There is a purpose-built bas_relief.json workflow already packaged inside the DepthAnythingV3 node pack — this is your fastest path.

Github

Github. ComfyUI-DepthAnythingV3/workflows/bas_relief.json at main · PozzettiAndrea/ComfyUI-DepthAnythingV3 · GitHub

📦 Step 1 — Install These Two Custom Node Packs

Open ComfyUI Manager → Install via Git URL and add both:

Node Pack

GitHub URL

Purpose

ComfyUI-DepthAnythingV3

https://github.com/PozzettiAndrea/ComfyUI-DepthAnythingV3

Generates depth map from your image

ComfyUI-Depth2Mesh

https://github.com/timothywong731/ComfyUI-Depth2Mesh

Converts depth map → watertight STL

Windows users: If fast-simplification fails to install, grab the latest Visual C++ Redistributables first.

Github

Github. GitHub - timothywong731/ComfyUI-Depth2Mesh: Convert 2D depthmap (heightmap) to 3D mesh for CNC milling · GitHub

📥 Step 2 — Load the Built-in Bas-Relief Workflow

After installing ComfyUI-DepthAnythingV3, the workflow file is at:

Code

Copy

ComfyUI/custom_nodes/ComfyUI-DepthAnythingV3/workflows/bas_relief.json

Load it via ComfyUI → Workflows → Open and drag the .json in.

Github

Github. ComfyUI-DepthAnythingV3/workflows/bas_relief.json at main · PozzettiAndrea/ComfyUI-DepthAnythingV3 · GitHub

🔗 Step 3 — Node Chain (What the Workflow Does)

Code

Copy

LoadImage  ──►  DepthAnything_V3  ──►  Depth Map to Mesh  ──►  Simplify Mesh  ──►  Save Mesh as STL

                     │                        ▲

               (grayscale depth)     width_mm / height_mm

                     │                   depth_mm / power

               PreviewImage

Key node settings for your badge:

Node

Setting

Recommended Value

DepthAnything_V3

model

DA3MONO-Large for geometric accuracy

Depth Map to Mesh

width_mm

60 (or your target diameter in mm)

Depth Map to Mesh

depth_mm

6.35 (= 0.25 inches)

Depth Map to Mesh

power

1.2–1.5 (sharpens peaks on emblems)

Simplify Mesh

target_face_count

100,000 (safe for Bambu Studio)

💡 Tips for Black & White Emblem Images

Since your Rocco Laurie and NYPD images are already high-contrast grayscale, you get an advantage:

• Skip DepthAnythingV3 entirely — feed your heightmap PNG directly into the Depth Map to Mesh node. White = raised, black = flat base. This gives you exact control vs. AI-estimated depth.

• Set power > 1.0 (try 1.5–2.0) to exaggerate the relief and make engraved detail pop.

• Use the alpha channel of your PNG to mask a circular outline — only pixels inside the circle will be meshed, giving you a shaped coin/medallion instead of a rectangle.

• Github

• Github. ComfyUI-Depth2Mesh/README.md at main · timothywong731/ComfyUI-Depth2Mesh · GitHub

⚡ Shortcut — Skip ComfyUI Entirely

If you just want the STL right now without setting up ComfyUI, the heightmap PNG I already generated for the Rocco Laurie emblem is ready to go. I can feed it directly into the Python mesh pipeline and generate the STL — want me to do that now?