Thirteen document tools where the file never leaves the browser
Staff were uploading confidential documents to free online PDF tools. Everyone understood it was a problem, and everyone kept doing it, because the alternative was doing the job by hand.
The habit nobody could police
Confidential documents were going to third-party tools, which was a privacy and compliance risk the client could not accept. Banning the tools does not work. People have a document to split and a deadline, and the free site is one search away.
What we built
Thirteen PDF and image tools that run entirely in the browser. The file is processed on the machine it is already on and never hits a server, which turns the compliance question into a non-question rather than a policy. There is a canvas-based PDF editor with inline editing, and a private AI assistant running on a self-hosted Ollama model.
Why in-browser rather than a private server
A private upload endpoint would have been easier to build and would still have created a copy of every confidential document in a second place. Processing in the browser means there is nothing to secure, nothing to retain, and nothing to explain to an auditor. The safest data is the data that was never transmitted.
Stack
React and TypeScript on Vite, with pdf-lib doing the document work client-side. Express for the small amount that needs a server, and a self-hosted Ollama model behind the assistant.
Want the same thing looked at in your business?
We map the process, find the part worth automating first, and give you a written scope with the phases separated. No obligation, and you keep the plan.