Klarity
What it is
Klarity is a free desktop tool for restoring images and videos locally and offline on Linux or Windows. Its processing options cover denoising, deblurring, upscaling, cleaning, and full restoration, as well as frame generation and related clean-frame and full-frame modes. It supports JPG, JPEG, PNG, BMP, TIFF, TIF, and WebP images, plus MP4, AVI, MOV, MKV, WebM, FLV, WMV, and M4V videos. Video processing requires FFmpeg. Users can work through a graphical interface or interactive and direct command-line modes. Heavy mode downloads about 888 MB of models, while Lite downloads about 204 MB and is described as 20x faster. CPU processing works without a GPU; the listed minimum requirements are a 2-core CPU, 4 GB RAM, and 2 GB storage. Models download on first use. Klarity uses NAFNet for denoising and deblurring, Real-HAT-GAN or Real-ESRGAN for upscaling, and RIFE for frame generation. Its license is MIT, with separate licenses for model components.
For and against
For
- Works locally and offline
- Processes images and videos
- CPU processing works without a GPU
- Offers graphical and command-line interfaces
Against
- Video processing requires FFmpeg
- AI models download on first use
- Heavy mode downloads about 888 MB of models
The verdict
Klarity offers several restoration modes and broad image and video format support in a free desktop package. Account for the model downloads and FFmpeg requirement for video work.
Who it suits
Klarity may suit creators and editors who want local image or video restoration without subscription fees. It also fits users who prefer a graphical interface or command-line workflow.
Plans and prices
Klarity
Free, local and offline image/video restoration tool
Read 2 Oct 2026 · github.com
On the record
- Download availability
- The latest release page lists CPU-only desktop binaries for Linux x86_64 and Windows x86_64.github.com · read 2 Oct 2026
- Integrations
- It uses NAFNet for denoising and deblurring, Real-HAT-GAN or Real-ESRGAN for upscaling, and RIFE for frame generation.github.com · read 2 Oct 2026
- Interface
- The tool provides a graphical interface and command-line modes, including interactive and direct CLI commands.github.com · read 2 Oct 2026
- License
- The repository lists Klarity under the MIT license and gives separate licenses for its model components.github.com · read 2 Oct 2026
- Model modes
- Heavy mode downloads about 888 MB of models for best quality, while Lite downloads about 204 MB and is described as 20x faster.github.com · read 2 Oct 2026
- Offline use
- The maker describes Klarity as local and offline; its release notes say AI models are downloaded on first use.github.com · read 2 Oct 2026
- Processing modes
- It offers denoise, deblur, upscale, clean, full, frame generation, clean-frame generation, and full-frame generation modes.github.com · read 2 Oct 2026
- Purpose
- Klarity is a local, offline AI tool for restoring images and videos through denoising, deblurring, upscaling, and frame generation.github.com · read 2 Oct 2026
- Support and audience
- The README says Klarity is built for creators, editors, and anyone who needs to enhance visual content without subscription fees.github.com · read 2 Oct 2026
- Supported formats
- It supports JPG, JPEG, PNG, BMP, TIFF, TIF, and WebP images, plus MP4, AVI, MOV, MKV, WebM, FLV, WMV, and M4V videos.github.com · read 2 Oct 2026
- System requirements
- The README lists minimum requirements of a 2-core CPU, 4 GB RAM, and 2 GB storage; CPU processing works without a GPU.github.com · read 2 Oct 2026
- Video requirement
- FFmpeg is required for video processing.github.com · read 2 Oct 2026
