Most stories on this site follow AI out of a lab and into a hospital, a farm, or a power plant. This one starts closer to home. Gregory Way, who runs appliedscience.com, has had strokes. They left him with vascular dementia and with trouble speaking. Talking, the thing most people do without thinking, became work. So he built a mobile app to give that back, for himself and for people in the same spot. He calls it Conversation Bridge for now. (That is a working name while he checks that it is available, so it may change before launch.) The applied point is not a research prototype. It is one stroke survivor taking AI tools anyone can rent and turning them into the thing he needed.
The gap he is trying to close is common. Stroke is the leading cause of aphasia, a disorder that affects the ability to speak and understand language, and the National Aphasia Association estimates that about a third of stroke survivors have it. Vascular dementia, the condition Gregory lives with, is caused by reduced blood flow to the brain, and a stroke is one of the ways that happens. Those are the medical labels. The everyday version is simpler: someone asks you a question, you know what you want to say, and the words do not come out fast enough, or clearly enough, to keep up.
Gregory noticed that deaf people face the same gap from the other side. For him, the hard part is speaking. For someone who is deaf, it is hearing. Either way, a normal back-and-forth conversation breaks down, and the person on the other end often does not know how to fix it. So Conversation Bridge is built for both groups.
Here is how it works. You tap to listen. The app takes what the other person says, turns it into text, and shows it on the screen. Then it offers optional quick-response buttons, or you can type your own reply. When you are ready, the app speaks your reply out loud. The other person hears your answer, and you are understood again. That covers both directions. Someone who has trouble hearing can read what was said, and someone who has trouble talking can still answer out loud.
The parts under the hood are off-the-shelf. Gregory built the app with Astra, an app builder, and the listening side runs on OpenAI's Realtime API, a service designed to handle live speech, which turns the other person's words into text. A few years ago, live speech-to-text, a mobile interface, and a voice that speaks for you would each have been their own engineering project. Now they are building blocks a single person can put together. That is the same timeline compression this site keeps covering in drug discovery and materials science, applied at the scale of one person.
It also changes who gets to build. Assistive technology has usually been designed by companies for users they study from the outside. Here the builder is the user. Gregory knows which moments in a conversation go wrong because he lives them. The quick-response buttons, the typed fallback, and the voice that speaks for you come from that experience, not from a focus group.
Conversation Bridge is early. It is set to arrive on Google Play this Thursday, October 8. Until then, a pre-release page on Gregory's Code and Host site has more about the app and will link to the store listing as soon as it is live: https://code-and-host.com/apps/conversationbridge.html.
For appliedscience.com, the hook is simple. The big AI breakthroughs get the headlines, but the quiet change is who can now use them. A stroke survivor who has trouble speaking took a speech-to-text service and an app builder and made a tool so he, and people like him, and deaf people too, can hold a conversation again. The science that already shipped is the building blocks. The application is his.