No. DepthForge works completely offline. All required AI models and runtime components are bundled with the add-on. No cloud services, API keys, accounts, or subscriptions are required.

No. DepthForge uses CPU-based AI inference, so a dedicated GPU is not required. Processing speed depends on your CPU, input resolution, and available memory.

Yes. DepthForge generates depth sequences from videos or image sequences and uses them to drive animated relief geometry in Blender. The original footage can be displayed on the relief as a material.

DepthForge creates depth-driven 2.5D relief surfaces, not complete 360-degree object reconstructions. AI-estimated depth is relative and does not represent exact real-world distances.

DepthForge provides separate packages for Blender 4.2–5.0 and Blender 5.1+, covering Windows x64, macOS Apple Silicon, macOS Intel, and Linux x64. Current development testing has been completed on Linux with Blender 4.2.0 and 5.2.2; other platform builds are pending final validation.

Yes. Every relief uses a procedural Geometry Nodes setup. You can adjust Height, Size, Resolution, Gamma, smoothing, depth cutoff, thickness, and more without regenerating the AI depth map.

Yes. Image depth results can be exported as 16-bit PNG heightmaps. Video processing produces 16-bit PNG depth sequences. Convert to Mesh also lets you turn the current relief into standard Blender mesh geometry.

Video inputs include MP4, MOV, MKV, AVI, WebM, M4V, MPG, GIF, and MXF. Supported numbered image sequences include PNG, JPG, TIFF, EXR, BMP, WebP, and TGA. Actual video decoding depends on the file's codec and platform.

Memory requirements depend on resolution and clip settings. High-detail Video Depth Anything processing may require approximately 4–5 GB of available RAM. Lowering Detail can significantly reduce memory use and processing time.