How I built in-browser PDF editing without uploading files

I recently worked on building a browser-based PDF tool where users can edit PDFs without uploading them to a server. The main goal was: - Keep everything client-side (privacy-first) - Support common operations like merge, split, reorder, and basic editing - Avoid heavy backend processing Here’s how I approached it: Rendering PDFs in the browser - Used a PDF rendering library to display pages on canvas - Extracted text layer separately for interaction Editing approach - Instead of modifying the original PDF directly, I used an overlay-based editing system - When a user clicks text: • Show an input field on top of it • Capture changes The tricky part: updating the PDF - PDFs don’t support direct text editing like HTML - So I had to: • Cover the original text area • Draw new text at the same position Handling layout issues - Text grouping vs individual elements - Position accuracy (X/Y alignment) - Font mismatches and spacing differences Performance considerations - Avoid re-rendering ent