PixelForge AI

PixelForge automates a two-stage image enhancement pipeline on Windows: AI upscaling followed by smart compression, run as a single batch job across entire folders.

Instead of manually running an upscaler on each image and then a separate compressor, PixelForge queues folders and individual images, processes them through both stages automatically, and tracks progress with per-image counters, elapsed time, and ETA.

Key features:

• Three pipeline modes: Upscale + Compress, Upscale only, or Compress only
• 7 bundled AI models (Standard, Lite, Ultra Sharp, Remacri, UltraMix Balanced, Digital Art, High Fidelity) covering photography, illustration, and fast/low-VRAM use cases
• GPU-accelerated inference via Vulkan (NVIDIA, AMD, Intel), with automatic dedicated-GPU priority
• Input queue supporting multiple folders and hand-picked images in a single run, with recursive scanning and folder structure preservation
• Pause, resume, and cancel mid-batch
• Before/after comparison slider for reviewing results
• Custom output naming templates ({name}, {model}, {scale})
• Desktop notifications on batch completion
• 100% offline and private — no cloud, no accounts, no telemetry

PixelForge is an automation layer built on the open-source upscayl-ncnn engine (AGPL-3.0) and Caesium CLT (GPL-3.0), adding the batch queueing, GPU management, and pipeline orchestration neither tool provides on its own.