V
Video2X
Description
Video2X is an open framework for machine-learning video super resolution and frame interpolation. It offers two modes: filtering, which upscales the picture, and interpolation, which adds frames. Inside it runs Anime4K v4, Real-ESRGAN, Real-CUGAN and RIFE through ncnn and Vulkan, and it also accepts custom GLSL shaders. It is installed as a Windows installer, as Linux packages or an AppImage, or as a container image, and there is a Google Colab notebook. The GPU must support Vulkan; the authors list the minimum generations of NVIDIA, AMD and Intel cards. The code is under the GNU AGPL v3 license, a copyleft license that has conditions for reuse, so read the license before commercial use.
Pros and cons
Pros
- Upscaling and frame interpolation in one tool
- Several engines to choose from, including Real-ESRGAN and RIFE
- Windows installer, Linux packages and container images
- Runs on any GPU with Vulkan support, not only NVIDIA
Cons
- AGPL v3 is a copyleft license with conditions, check it before commercial use
- Needs a Vulkan-capable GPU and an AVX2 processor for the ready builds
- Setup is more technical than a regular app
- Latest stable release dates from early 2025
Key facts
- License
- GNU AGPL v3
- Commercial use
- Allowed with restrictions
- Hardware requirements
- GPU with Vulkan support; precompiled builds need a CPU with AVX2
- Developer
- K4YT3X and contributors
- Checked on
- Oct 21, 2026
Editorial score
How strongly the AI Expertise editors recommend this open model for each of its tasks right now, on a 0–10 scale. The score combines output quality, capabilities, ease of launch, license openness and freshness.