Client-side Image & PDF Compression

Compress images and PDFs
without sacrificing quality

Batch compress visual files and partition multi-page PDFs under 5MB or 10MB targets. All processing happens in your browser.

How It Works

A streamlined 3-step workflow for high-efficiency compression

1

Select or Drop Files

Drag and drop individual images or multi-page PDF documents. Files are read directly in browser memory without sending a single byte to any remote server.

2

Configure Targets & Filters

Choose your output format (WebP, JPEG, PNG, AVIF), adjust quality, or pick an automated target size budget like < 5MB or < 10MB.

3

Convert & Download

Watch page-by-page execution in partitioned windows. Download the re-assembled compressed PDF directly or extract a ZIP archive containing individual pages.

Under the Hood

Image & PDF Processing Pipeline

Zero-server, low-memory client architecture designed for heavy files

Stage 01

Ingestion & Detection

Differentiates images from PDF documents using Web Streams. Extracts page count and layout geometry.

Stage 02

Partitioned Rasterization

5-page chunk windows with automatic garbage collection yields to keep memory flat under 200MB even on 300-page docs.

Stage 03

Adaptive Compression

Quantization tables auto-throttle per page based on remaining budget (<5MB / <10MB). Optional grayscale filter.

Stage 04

Assembly & Packaging

Pure client PDF binary assembler stitches pages into a unified document or JSZip bundles organized folders.

Image Processing Path
File → HTMLCanvas → ImageData → Canvas/WASM Encode → Single/ZIP Download

Multi-threaded parallel queue (2 concurrent workers) with real-time compression ratio calculation.

PDF Processing Path
PDF ArrayBuffer → Partition Loop (5 pgs) → Canvas Render → Cleanup → PDF Assembler

Sequential chunk rendering with explicit page.cleanup() and zeroed canvas backing stores to avoid memory leaks.

Engineered for quality and privacy

Lightweight, fast, and completely contained in your browser

Local & Private

Your files are never uploaded to a cloud server. Memory deallocates directly in your browser.

Partitioned PDF Engine

Rasterizes and compiles heavy documents in 5-page partitions to eliminate memory spikes.

Target Budgeting

Adaptive quantization guarantees outputs conform to strict < 5MB or < 10MB limits.