A long-running open framework for AI video upscaling and frame interpolation, with several backends in one tool.
Worth knowing
Price: Open source (AGPL-3.0)
License: AGPL-3.0
Status: sourced. We have not run this ourselves yet (no "Tested" badge).
Quality bar (computed by code): 4/5 (primary source ✓, reproducible ✗, new ✓, example output ✓, safe ✓)
Get Atlas Dispatch
The best new AI video workflows and prompts, weekly.
Weekly: 5 new prompts and workflows. Unsubscribe anytime.
From the source
Setup. Windows installer (Qt6 GUI), Linux packages (AUR plus a universal AppImage), container images on GitHub Container Registry for Linux and macOS, or a free Google Colab notebook (T4/L4/A100, up to 12 hours a session). It needs a CPU with AVX2 and a Vulkan-capable GPU (NVIDIA Kepler, AMD GCN 1.0 or Intel HD 4000 or newer). AGPL-3.0.
Inputs. A video file, plus a choice of filtering (upscaling) or frame interpolation. Upscalers: Anime4K v4 and custom MPV GLSL shaders, Real-ESRGAN and Real-CUGAN. Interpolation: RIFE. All run via ncnn and Vulkan.
What the source says about results. Version 6 is a C/C++ rewrite. The author says it's much faster than the 5.0 beta, gives better output quality and needs no extra disk space during processing beyond the final file.
Connect Claude, ChatGPT, Cursor, Gemini or Grok Bot to Atlas, and your agent can plan shots, pick models and price them from the data on this site. Free. No sign-up or API key.
▶ Play demoYouTube · loads on click