Skip to content

Visual Enhancer Documentation

Visual Enhancer combines NVIDIA RTX neural pipelines, DirectX 12 optical-flow effects, ReShade shaders, PyTorch AI restoration models, and 16-bit color grading in a single portable Windows desktop app and scriptable CLI (VE_CLI.exe).

Visual Enhancer Neural Rendering workspace showing Split preview of an ancient Egyptian temple harbor vista, Batch Queue drawer, and DLSS Neural Rendering inspector controls

Input: Comparison between original ancient Egyptian temple harbor panorama and DLSS Neural Rendering Style 0 outputDLSS Neural Rendering: Comparison between original ancient Egyptian temple harbor panorama and DLSS Neural Rendering Style 0 output
Interactive preview rendered with DLSS Neural Rendering (Style 0, NR Intensity 1.05, 1 Pass, Local Tone 0.85, Local Structure 0.90).

Looking for live motion comparisons? Explore interactive before/after video demos for Frame Generation (24 → 120 FPS), Temporal AI Video Upscaling, and Multi-Stage Video Restoration, or browse full stacks under Recommended Video Recipes.

  1. Load media — Drag and drop images or videos, paste files or image URLs with Ctrl+V, or queue entire folders in Neural Rendering (Image or Video mode).
  2. Build your stage stack — Enable and drag-reorder Pipeline cards in the right inspector, or leave every card off for a clean format conversion pass-through.
  3. Preview before rendering — Compare Split, 2-Up, or Output views, drag across the video timeline, or step frame by frame with Left / Right.
  4. Render or automate — Export locally with Render Image(s) / Render Video(s), or export a .json preset for headless batch jobs via VE_CLI.exe.

System Requirements

Check NVIDIA RTX, Windows HAGS, DirectX 12, CUDA / XPU / ROCm NeuroFrame Runtime, and resolution ceilings in System Requirements.

Export & Codecs

Compare 8/16-bit PNG, TIFF, JPEG XL, WebP, AVIF, NVENC, AV1, VP9, ProRes, and FFV1 Lossless RGB in Export & Formats.