apex-compute/ANR

What ANR does

Apex Neural Rendering (ANR) is a game rendering enhancement suite from Apex Compute. The currently available component, ANR FG, adds frame generation to supported games; super resolution (ANR SR) and ray reconstruction (ANR RR) are listed as coming soon. Installation is game-specific: download a package, extract it, and run the installer against the folder containing the game's .exe. No source code is published yet.

The project targets players on Windows 10/11 or Linux/SteamOS with Proton. It supports NVIDIA RTX 20–50 and GTX 16 series, AMD RX 5000–9000 series and Steam Deck, and Intel Arc A/B series.

Quick start

1. Open the Releases page and choose anr-fg-full for games that do not use OptiScaler, or anr-fg-addon for games that already do.
2. Close the game, extract the package, and run the installer with the folder holding the game's .exe:
   • Windows: install.cmd
   • Linux/Steam Deck: sh install.sh "<folder>", then set the launch options to the command it prints, such as WINEDLLOVERRIDES="dxgi=n,b" %command% (use winmm=n,b for Vulkan games).
3. Launch the game and enable frame generation, using the NVIDIA option when the game presents several choices.

To remove it, run uninstall.cmd or sh uninstall.sh with the same game folder.

Constraints and diagnostics

ANR FG is in beta, and only the packaged installer is available; the README states that source code is coming later. The OptiScaler dependency is GPL-3.0, and its source is attached to each release. Results and problems belong in the issue tracker: performance reports should include the test kit's result file, and bug reports should include the game, GPU, driver, OS, settings, and the latest nr_fg-*.log from the game's OptiScaler/custom/Apex Neural Rendering FG folder.

HotRepo's take

The repository entered the all-language new list on 2026-11-12 with 32 stars gained in roughly one hour, reaching 141 stars and 2 forks by the recorded snapshot. That is early traction for a beta game addon with no public source or declared license. The release-based distribution and lack of implementation detail are the main things to review before installing.