I built a $0 infrastructure Table-to-Excel tool using client-side WASM and Next.js. Roast my architecture.

I got tired of clunky OCR tools that charge a subscription or, worse, upload sensitive financial data to the cloud.

So, I built TABLEXTRACT. It runs entirely in your browser using Next.js and Tesseract.js (WASM). Since all the OCR and Excel file generation happens locally, your data never leaves your machine, and my server cost is exactly $0.

I'm validating the idea before buying a proper domain, so it's currently on a free Cloudflare staging URL.

👉 *Try it here: https://tablextract.itsanssoff.workers.dev/

Please throw your messiest screenshots at it. I’d love to know what breaks and get your thoughts on the performance of client-side OCR!