About
One person, building an assistant that stays on your machine
AssistantX is a voice assistant for your desktop. It listens, answers, and does things for you — and the work happens on your own hardware, not in someone else's datacentre.
Why it exists
Most voice assistants are a microphone pointed at a company's servers. You talk, your words leave your computer, and what happens to them afterwards is not really your business. AssistantX started as an objection to that arrangement: the assistant should live where your files already live.
Not a chatbot. An operator.
A chatbot writes text back at you. AssistantX is built to act — open an application, change the volume, look something up, take on a longer job in the background and tell you when it is finished. Answering is the easy half; doing is the point.
It runs on your hardware
Speech recognition, the model that thinks, and the voice that answers all run in the desktop app itself. It is deliberately built to fit an ordinary graphics card rather than a rack of them — a machine you might actually own, not one you rent by the hour.
You just say the word
Nothing to click. A wake word starts it, the assistant confirms it heard you, and from there you simply talk. Knowing when a sentence has ended, not talking over you, hearing someone speaking quietly at midnight — these sound trivial and are most of the work.
Made by one person
There is no company here, no team, no investors. One person — Wiktor — designs it, builds it and decides what ships. That shapes the product more than any roadmap: it moves at the pace of one person, changes direction when something genuinely does not work, and does not grow features to fill a slide.
Where it is now
Early, and honest about it. The roadmap page shows the actual state of things rather than a wish list — it is the best place to see what works today and what is still being built.
Built in Poland. Still being built.