For everyone who lives in calls
“Sorry — could you repeat
the question?”
You drifted for ninety seconds. They said your name. Hearname sits in the tray, listens to the call on your own machine, and pulses the screen the second it is you — with the last twenty seconds as text.
…so the migration slipped a week, which is fine, but Mikhail , you had the numbers on retention — what did they say?
Pulsed 0.4s after it was said. Nothing recorded, nothing uploaded.
The maths of a working day
You are not going to stop drifting. Nobody does.
21h
a week in meetings, for the average manager
4 in 5
admit to doing other work during calls
2.4s
the silence after your name that everyone remembers
Most of a long call genuinely does not concern you — that is why you answer email through it. The problem is not attention, it is that nothing tells you when the twenty seconds that do concern you have started.
How it works
It hears what your speakers hear
System audio, straight from the machine. No bot joins the call, nobody sees a new participant, and it works the same in Zoom, Meet, Teams, Discord or a browser tab.
It listens on your machine
Speech recognition runs locally. The free edition never sends audio anywhere — there is no upload to switch off, because there is no upload.
It pulses when it is you
Your name, its diminutives, and the ways people get it wrong. The screen edge pulses coral and the last twenty seconds are on screen as text, so you can answer instead of asking them to repeat it.
What makes it liveable
It knows how your name gets mangled
You add “Mikhail” and it also catches Michael, Mikhael, Misha, and the version your American colleague has been saying for two years. Fuzzy matching by sound, not by spelling — which is the whole difficulty.
Nothing is recorded unless you ask
Audio passes through a few seconds of memory and is gone. Turn on history and it is written to a local database on your disk — never to us, on any plan.
It is not a meeting bot
No calendar access, no third participant, no “this meeting is being recorded” banner. Legally and socially those are a different thing, and this is deliberately not it.
Small enough to forget about
A native app in the tray, single-digit CPU, no browser tab, no account required to run the free edition. It should cost you nothing to leave running all day.
Pricing
The part that saves you costs nothing.
Local detection is free and stays free — the core is open source. Pro is for accuracy and for finding, three weeks later, the meeting where your name came up.
Free
$0 /mo
Local detection, forever, on one machine.
- — Local speech recognition
- — Name & alias matching
- — Screen pulse
- — Open source core
Pro
$6 /mo
You want it accurate and you want the last minute back.
- — Higher-accuracy cloud transcription
- — Searchable history across calls
- — Who said it (speaker labels)
- — Multi-language calls
Prices at launch. Early-access members get their first month of Pro free.
Questions
Is this legal? Am I recording people? +
By default nothing is recorded — audio lives for a few seconds in memory and is discarded. If you turn on history, you are storing a transcript locally and consent rules where you live apply to you, exactly as they would for any note you take. We put that in plain words in the terms rather than burying it.
Does it join the meeting? +
No. It reads the sound your machine is already playing. Other participants see nothing, because from the call’s point of view nothing happened.
Does my audio leave my computer? +
Not on the free plan — recognition runs locally, on your CPU. On Pro you can switch to cloud transcription for accuracy; then audio goes to the transcription provider only while that is on, is not retained afterwards, and is not used to train models.
Which languages? +
The local model handles the major languages out of the box, including calls where people switch mid-sentence — which is exactly when you stop following and most need this.
What if it mishears? +
It sometimes will, in both directions. A false pulse costs you a glance; a missed one costs you the thing you were trying to avoid — so it is tuned to err towards pulsing. You can tighten it.
EARLY ACCESS
Stop missing your turn.
We let people in a few at a time so we can fix what breaks. Tell us how your calls go and you go in the next batch.
The two questions below are not a formality. Everyone who fills them in gets a personal reply from me within a day.