Multi-engine browser OCR with visual results. Choose between fast mobile models (PPU) or accurate server models (OnnxOCR). Features Tesseract fallback, visual bounding boxes, 100% client-side processing, PWA support, and intelligent model caching.
