| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| Visual.Enhancer.v21.0.zip | 2026-11-02 | 824.0 MB | |
| README.md | 2026-11-02 | 1.2 kB | |
| Visual Enhancer v21.0 source code.tar.gz | 2026-11-02 | 9.7 MB | |
| Visual Enhancer v21.0 source code.zip | 2026-11-02 | 11.5 MB | |
| Totals: 4 Items | 845.2 MB | 0 | |
Changelog
- Added RazorNext, a new state-of-the-art AI image upscaling model for enhancing images at 1×, 2×, 4×, 6×, or 8× scale. It works best with images of at least 720 pixels. Adjust GPU Memory with Auto, Unlimited, or a fixed limit to suit your hardware.
- Improved model downloads, with more informative progress and smarter retries that respond to server delays and temporary connection problems.
- Improved processing time estimates, giving image and video batches a clearer indication of remaining time while avoiding a premature countdown to zero.
- Improved ReShade compatibility, letting supported techniques continue when others cannot run and explaining when depth effects are skipped because depth input is disabled.
- Refined video preview and media handling, including seeking improvements and additional coverage for ProRes imports.
- Improved AI model loading and GPU caching across the upscaling engines, with updates to Windows ML compilation and runtime handling.
- Refreshed interface translations and model management, including clearer download states and the shorter GPU Memory label.