This happens if a *different* Python package (from another addon, or a previous manual install) overwrote the AI library versions this addon needs. Fix: open a terminal, point it at Blender's Python (see the README for the exact path), and run: ``` pip uninstall torch transformers -y pip install torch==2.1.0 --index-url https://download.pytorch.org/whl/cpu pip install transformers==4.46.3 ``` This reinstalls the exact versions verified to work together. (Newer Blender releases with a newer bundled Python may not have this exact pinned version available — in that case the addon automatically falls back to the latest compatible versions, no manual steps needed.)

This usually means the depth map's contrast is concentrated in a narrow range. Use the **"Auto Contrast"** button in "Current Mesh Settings" — it analyzes the actual depth data (ignoring any fully-transparent background) and stretches the black/white points automatically. You can also fine-tune manually with the "Black Point" / "White Point" sliders.

This is almost always a missing Microsoft Visual C++ Redistributable, not an addon bug. Install the [VC++ Redistributable x64](https://aka.ms/vs/17/release/vc_redist.x64.exe), restart your PC, and try again.

Yes, 100%. The models, depth maps, and any content you generate with this addon are entirely yours — no royalties, no attribution required back to us.

Currently tested and supported on Windows only. Mac/Linux support may come in a future update depending on demand.

No. The addon auto-detects NVIDIA GPUs and uses CUDA acceleration automatically if available, but it also works fine on CPU — just slower per generation.

No per-generation cost, no images uploaded to a third-party server, no internet dependency after setup. Your photos never leave your machine.

No. The addon auto-detects NVIDIA GPUs and uses CUDA acceleration automatically if available, but it also works fine on CPU — just slower per generation.