I got tired of copy-pasting tables from PDF invoices into spreadsheets. The columns always break, numbers merge with text, and you end up spending more time fixing the formatting than the actual work.

So I built a tool that uses AI to extract every structured data block from a PDF — line items, totals, headers, metadata — and lets you download them as clean CSV.

How it works: - PDF is parsed client-side (your file never leaves your browser) - Text layout is reconstructed using positional data from the PDF - AI identifies every table and structured block - You select which blocks to export and download as CSV

Tech stack: Next.js, Anthropic API (Claude), Stripe, Vercel. PDF parsing is done with pdf.js in the browser.

Free for PDFs up to 3 pages. Feedback welcome — especially on extraction accuracy for different PDF formats.