Source-available • Windows first • CUDA ready

Advanced 2D to 3D conversion for cinema and VR

VisionDepth3D transforms flat video into high-quality stereo 3D using AI depth estimation and proprietary stereo rendering tools. Designed for creators who need precise control, stable depth, and professional export options.

16k+ downloads 204 GitHub stars Python 3.13
VisionDepth3D
VisionDepth3D 3D Generator
VisionDepth3D Frame Tools
VisionDepth3D Depth Blender
Hotkeys
Hotkeys
VisionDepth3D Logo

Built-in tools

A complete workflow for turning 2D footage into refined stereo 3D. Estimate depth, preview results, enhance quality, and export in formats suited for modern displays and headsets.

Depth Estimation

Support for 25+ depth models, including DPT, ZoeDepth, Depth Anything V2, Marigold, and more.

Upscaling

Integrated Real-ESRGAN upscaling for sharper final output, with support for custom models.

Live 3D Preview

Preview stereo output in real time on external displays or desktop sources.

Frame Interpolation

RIFE-based interpolation for smoother playback and higher frame rate delivery.

GPU Acceleration

CUDA and ONNX Runtime support, with TensorRT-friendly workflows for advanced users.

Batch Rendering

Process full folders or matched video and depth-map pairs with progress tracking.

Depth Blender

Combine multiple depth maps to improve scene separation and perceived depth quality.

Audio Tools

Attach, offset, and manage audio streams while maintaining sync during export.

Why creators use VisionDepth3D

Core rendering features focused on depth stability, visual comfort, and higher quality stereo output.

Per-Pixel Parallax

Depth-driven stereo shifting enables precise image separation with minimal distortion.

Zero Parallax Control

Keep key subjects grounded naturally within the stereo window for more comfortable viewing.

Edge-Aware Masking

Reduces ghosting and edge artifacts through smarter shift protection and pixel recovery.

Scene-Adaptive Depth Control

Smooths abrupt stereo changes across cuts and motion-heavy sequences.

Floating Window Tools

Adjust windowing to reduce edge violations and improve overall viewing comfort.

Flexible Output Formats

Export in Half SBS, Full SBS, anaglyph, and interlaced formats for a wide range of displays.

The VisionDepth3D Method

A proprietary hybrid workflow for real-time and offline 2D to 3D conversion, developed for cinematic presentation and immersive VR viewing.

  • Depth-weighted continuous parallax shifting
  • Subject-aware zero parallax tracking
  • Edge-aware shift masking to reduce ghosting
  • Scene-aware parallax dampening
  • CUDA-accelerated stereo warping
  • Depth-of-field simulation with pixel healing

The method is documented for reference and educational review. Reuse, redistribution, or derivative publication of the method text is not permitted.

Download

Download the Windows build and start producing stereo 3D locally. VisionDepth3D is optimized for Windows systems, with NVIDIA GPUs recommended for the best performance.

  • Can run on CPU or GPU, with CUDA acceleration recommended for faster processing
  • Source-available on GitHub under a personal, non-commercial license
  • No telemetry or cloud rendering. Your files stay on your system

Formats supported

  • Half Side by Side
  • Full Side by Side
  • VR180 equirect TB and SBS
  • Anaglyph
  • Interlaced

Minimum requirements

  • Windows 10 or 11
  • 8 GB RAM minimum, 16 GB recommended
  • NVIDIA GTX 10 series or better recommended

FAQ

Does it run on AMD or Apple GPUs?

VisionDepth3D can run on CPU across supported systems. GPU acceleration is currently focused on NVIDIA through CUDA and ONNX workflows. AMD users may be able to use alternative backends where available, while Apple Silicon support is planned for future development.

Can I use my own depth model?

Yes. Repository users can add compatible model weights to the appropriate folder and select them from within the application. Both PyTorch and ONNX-based workflows are supported, depending on the model.

What license does the app use?

VisionDepth3D is available for personal, non-commercial use under the VisionDepth3D License Agreement. Commercial, business, or revenue-generating use requires written permission from the licensor. Third-party components remain subject to their own licenses. See the LICENSE for full terms or email redsky90@gmail.com for commercial licensing.

Where can I get help?

Support is available through GitHub issues, the community thread, and the project’s Itch.io page.

License

A brief overview of the licensing terms. Refer to the full license for complete details.

Usage Grant

Free for personal, non-commercial use.

Commercial Use

Commercial or revenue-generating use requires written permission.

Restrictions

Redistribution, resale, sublicensing, reverse engineering, and derivative works are prohibited under the license terms.

Ownership

All rights remain with the licensor, except where third-party components retain their own licensing terms.

Third-Party Components

Included frameworks, models, and external tools remain governed by their respective licenses.

Warranty

The software is provided as is, without warranty of any kind.

Read the full license: LICENSE on GitHub. Commercial licensing: redsky90@gmail.com.