REAL Video Enhancer

Summary

REAL Video Enhancer is a free application for processing video on Windows, Linux, and macOS. It provides frame interpolation, upscaling, decompression, and denoising, and supports batch processing. Interpolation models include RIFE, GMFSS, and IFRNet; an example use is changing video from 24 to 48 FPS. Upscaling options include SPAN, OpenProteus, AnimeJaNai, AniSD, and AnimeSR, while DeH264 handles decompression and DRUnet and DnCNN are listed for denoising. Scene-change detection is included to preserve sharp transitions, and the application previews its latest rendered frame. Compute options are TensorRT for NVIDIA RTX GPUs, PyTorch for CUDA 12.6- or ROCm 6.2-capable GPUs, and NCNN for Vulkan 1.3-capable GPUs. The listed minimum requirements include a dual-core x64 CPU, Vulkan 1.3-capable GPU, and 16 GB RAM; NCNN also requires 4 GB VRAM and 1 GB free storage. TensorRT optimization can fail with limited VRAM, for which CUDA or NCNN are suggested alternatives. The project is AGPL-3.0 licensed and links to Discord and GitHub issues for support. Its Windows binary may be detected as a trojan due to a PyInstaller false positive.

Who it is for

REAL Video Enhancer suits users who want interpolation, upscaling, decompression, or denoising on a supported desktop platform. Its GPU backend options and stated hardware requirements are important to check first.

Compared on video upscaling software

Free plan

Yesgithub.com

Frame interpolation

Yesgithub.com

Batch processing

Yesgithub.com

Facts

Purpose

REAL Video Enhancer provides video frame interpolation, upscaling, decompression, and denoising.github.com · 5 Oct 2026

Upscaling

The README lists multiple upscaling models, including SPAN, OpenProteus, AnimeJaNai, AniSD, and AnimeSR.github.com · 5 Oct 2026

Other processing

The project lists DeH264 decompression and DRUnet and DnCNN denoising models.github.com · 5 Oct 2026

Scene handling

Scene change detection is included to preserve sharp transitions, and the app previews the latest rendered frame.github.com · 5 Oct 2026

Compute backends

TensorRT supports NVIDIA RTX GPUs, PyTorch supports CUDA 12.6 and ROCm 6.2 capable GPUs, and NCNN supports Vulkan 1.3 capable GPUs.github.com · 5 Oct 2026

Integration

Discord Rich Presence is supported for the Discord system package and Discord Flatpak.github.com · 5 Oct 2026

Platforms

The project documents Windows, Linux, and macOS support, including Ubuntu 22.04+ and macOS 15+ ARM or x86 builds.github.com · 5 Oct 2026

Hardware requirements

The README gives minimum requirements of a dual-core x64 CPU, Vulkan 1.3-capable GPU, 4 GB VRAM for NCNN, 16 GB RAM, and 1 GB free storage for NCNN.github.com · 5 Oct 2026

Known constraint

The README says TensorRT optimization can fail with limited VRAM and suggests using CUDA or NCNN instead.github.com · 5 Oct 2026

Security notice

The project warns that the Windows binary may be detected as a trojan and attributes this to a PyInstaller false positive.github.com · 5 Oct 2026

License

The repository identifies its license as AGPL-3.0.github.com · 5 Oct 2026

Support

The project links to a Discord community and offers support through GitHub issues.github.com · 5 Oct 2026

Steam purchase

Steam lists a $2.99 donation DLC that contains no content and requires the base REAL Video Enhancer application.store.steampowered.com · 5 Oct 2026

Where it ranks on Everything Xiaomi

• Best Video Upscaling Software in 2026#4 of 40
• Best AI Video Upscalers in 2026#5 of 33

Sources

• github.com/TNTwise/REAL-Video-Enhancer· checked 5 Oct 2026
• store.steampowered.com/app/4382640/REAL_Video_Enhancer_Donatio· checked 5 Oct 2026