The problem with a single block of text
Phone-wide live captions and most transcription apps produce a running transcript. That works for a lecture or a phone call. It does not work for a dinner table, where three people can answer the same question and the joke only makes sense if you know who made it.
A name on every line
Akoe follows the voices in the room all the time, not just at the moment each line is written. Every line appears under the name of the person who said it, in their own color, with the newest at the bottom.
- Name the people you talk to most; the app remembers them.
- Up to four voices are told apart at once. Anyone beyond that appears as “Someone” instead of being given a wrong name.
- Your own voice is left out, so the screen shows only what you need to read.
- If you lose the thread, one tap shows a short summary of the last two minutes.
I’ll call her tomorrow before ten.
Dile que llevamos el postre.
Dad, are you coming with us on Sunday?
Honest about what it cannot do
Two people talking over each other, or someone far from the phone, are the hardest cases. Akoe does its best and shows what it heard rather than hiding it. Put the phone on the table with the bottom edge pointing at whoever is talking and it hears best.
Questions
How many people can it tell apart?
Up to four voices at once, with names. Anyone beyond that shows as “Someone”.
Do I have to teach it everyone’s voice?
No. It follows the voices as they talk. You add a name when you want one.
