I Built a File Converter That Handles 16 Formats in the Browser — No Upload Required

#javascript #webdev #showdev #productivity

A client once sent me a PDF invoice and asked for it as a CSV. Then sent a HEIC photo that needed to be PNG. Then an XML config that needed JSON. Three different tools, three different uploads, three different privacy policies to read.

I got fed up and built one tool that handles all of it locally.

The Stack

The core is surprisingly simple. Each conversion is an "engine" — a small JavaScript module that takes input and returns output:

PDF to JPG: Canvas-based page rendering + per-page export
HEIC to PNG: heic2any library, all client-side decoding
JSON ↔ CSV: PapaParse for CSV, native JSON.stringify/parse
XML ↔ JSON: DOMParser for XML parsing, zero dependencies
SVG to PNG: Canvas drawImage with crossOrigin handling
Markdown to HTML: marked.js, 2KB minified

16 engines total, all running in the browser at formlyapp.org.

What I Learned

HEIC is everywhere now. iPhone photos come as HEIC by default. Converting them requires the heic2any library which decodes the HEIF container in the browser. The first version crashed on files over 10MB — had to add chunked processing.
PDF rendering is inconsistent across browsers. Each browser renders PDFs slightly differently via Canvas. Firefox handles embedded fonts better than Chrome. Safari sometimes drops transparency. Testing across browsers took longer than writing the code.
File size limits are self-imposed. Most online converters cap at 25-50MB because they process server-side. Browser-local means the only limit is the user's RAM. On a 16GB machine, 200MB files process fine.