28 of our 30 PDF tools never touch a server. Here's how we built them

Most free PDF sites operate by having you upload your document, their server processes it, and you download the outcome. While suitable for a menu, this method falls short for sensitive documents like payslips, supplier contracts, or staff ID scans. When Regentflow developed PDF Studio, they established a rule: by default, the file remains on the device. Out of 30 tools, 28 adhere strictly to this principle. The remaining two offer an optional AI feature, which they disclose before execution.

The entire operation takes place within the browser utilizing four libraries: pdf-lib for writing, merging, splitting, and editing; @pdf-lib/fontkit for embedding fonts; pdf.js (pdfjs-dist) for page rendering and parsing; and SheetJS for transforming Excel files into PDFs. The browser accesses the file locally, performs computations in JavaScript on the page, and returns the result as a download. Consequently, there is no upload phase, eliminating any risk of data leakage, retention, or deletion on the server side.

Two exceptions to this rule are Redact PDF and PDF Editor. Redact PDF includes an optional AI function for scanning pages, while PDF Editor provides AI capabilities for reading and rewriting. These tools require manual intervention, sending an image of a single scanned page to their server and employing an AI model for processing. Users incur credits for these services and are charged a fee of A$1 per edited file (A$5 for ten files). All other components of these tools remain local, ensuring data privacy and security.

Regentflow, primarily a business-tools provider for cafes, bars, and restaurants, built this system to address user complaints about needing to fix PDFs elsewhere. Now, their users do not have to leave the platform to address PDF-related issues. To explore these tools, visit regentflow.io/pdf-tools. The question remains: which tool should they develop next?