TraceForge
Raster to Vector in Seconds.
GPU-accelerated vectorization engine combining neural upscaling (Real-ESRGAN) with precision tracing algorithms (Potrace, VTracer). 10+ presets optimized for logos, illustrations, and line art.
Neural Upscaling
Full vectorization is functional — Potrace, VTracer, all presets, and SVGO optimization work as expected. Real-ESRGAN neural upscaling requires GPU backend to be online.
● GPU not required for vectorizer
10+ Presets
Optimized for logos, icons, and illustrations
Dual Engines
Potrace + VTracer for any vectorization need
Open Source
Full source code available on GitHub
Clean Output
SVGO-optimized, production-ready SVGs
01Upload Image
Drag & drop or click to upload
PNG, JPG, WebP, GIF, BMP • Max 20MB
02Choose Preset
Pre-processing
Uses AI (rembg) to remove backgrounds before vectorization. Recommended for images with white or colored backgrounds.
Tips for Best Results
- •Use high-contrast images with clean edges
- •Enable “Remove Background” for images with white/colored backgrounds
- •Try “Logo Smooth” preset for most logos
- •Use “Geometric” for sharp-cornered designs
- •VTracer works better for multi-color images
- •Higher resolution input = better quality output
Technical Architecture
Tracing Engines
Potrace (classic bitmap tracing) + VTracer (modern multi-color vectorization)
Neural Upscaling
Real-ESRGAN 4x super-resolution with CUDA acceleration
Optimization
SVGO pipeline for production-ready, minified SVG output
Stack
Python/FastAPI backend, React frontend, Docker deployment