upscale-photo

Free open-source AI image & video upscaler. Upscale photos to 8K and videos to 4K locally with Real-ESRGAN + GFPGAN face restoration. A free Topaz Gigapixel / Video AI alternative for macOS, Windows & Linux — no cloud, no watermark, no subscription.

Images look blurry on a 4K/HiDPI screen over 100% display scaling, or on Retina? The browser upscales them with a cheap bilinear filter. This resamples them with a real Lanczos3 filter on the GPU instead. No third-party requests; nothing leaves your browser. Built for manga, comics, and webtoons, works anywhere.

Local free AI upscaling

AI-powered image upscaler — enhance photos up to 4x resolution, free, no signup, no watermark.

Modular, memory-aware AI image editing toolkit powered by SAM2, FLUX, OmniPaint, and Real-ESRGAN

open-source, browser-native image editor that combines traditional layer-based editing with ComfyUI integration and AI tooling.

基于 NVIDIA RTX VSR 神经网络的轻量 AI 图片放大 Windows 桌面工具，支持 X1/X2/X4 倍率，Tauri 2 构建。

Tessera is a open-source AI desktop image enhancer built for maps, text, and photos. Featuring a dual smart-math and AI pipeline, Tessera combines classical edge sharpening with line-art and photo models. Scale up to 16x with custom double-pass, TTA, and auto-polish modes for razor-sharp map labels or natural photo restoration.

PrivaScale is a mobile-first PWA that upscales images entirely in the browser. No server, no tracking, no cloud. It starts immediately with the bundled AI model — no setup screens.

Generate AI videos via OpenRouter and optionally upscale them with ffmpeg

AI image upscaler that enhances images 4× and exports JPEG files of at least 4 MB, with single-image and ZIP batch processing.

Single-file AI image upscaler CLI. Loads ESRGAN, Real-ESRGAN, DAT, HAT, SwinIR, SRFormer and friends via spandrel.

Self-host an app that lets you upscale images to 4k or 8k.

Image Upscaler

Upscale Studio is a Windows desktop application for batch image upscaling. It provides a PySide6 GUI and uses the upstream nunif/waifu2x PyTorch backend for the actual ML processing.

An open-source, lightweight alternative to Canvas built entirely in Pyside6.

A high-performance cross-platform C/C++ Super-Resolution image upscaling engine powered by ONNX Runtime.