Processed in Europe
The default recorded mode sends audio to European infrastructure for high accuracy transcription with speaker labels.
Meetings
Every meeting becomes decisions and tasks
One button captures both sides of the call, labels who said what, and hands you a summary, action items and a task board. No bot joins your call, because the app records on your own machine.
- No bot in your call
- Both sides recorded separately
- On-device mode available
- Export as text, Markdown, SRT or JSON
Your microphone and the system audio are captured as two separate tracks, with no bot joining the call.
You
Others
YouCan we commit to shipping the export by the 14th?
OthersOnly if the API review lands on Friday, otherwise it slips a week.
YouThen let's put the review on Friday morning and decide straight after.
OthersWorks for me. I'll send the endpoints tonight so there's time to read them.
Sync is yours to switch on. Off, everything stays on this computer. On, the transcript, summary and notes reach your other computers and any AI you connect. The audio is never synced either way.
Live now: You. Only that track carries signal.
Your microphone and the other side are separate tracks from the first second, so who said what is never a guess.
Meetings
How it works
01
Three ways to run a meeting, and you pick when you start
Live transcription for following along as people speak. Record with cloud transcription for maximum accuracy, processed in Europe. Record with on-device transcription for meetings that must never touch the cloud.
02
It captures both sides of the call
Your microphone and your computer's audio are recorded as separate tracks, so you and everyone else are labelled correctly from the first second. Works with any conferencing tool because it records the machine, not the meeting service.
03
Or capture a room
An in-person mode records a single microphone for meetings with no call at all, with a presentation captions option that shows live subtitles to an audience while you speak.
04
It knows who said what
Rename Speaker 1 to Alice and the summary uses her name. You tell it how many people were on the call, a one tap answer instead of an algorithm guessing and inventing phantom speakers. Echo cancellation strips the other person's voice out of your microphone track.
05
Talk turns into work
One click for a structured protocol, a summary, action items, a follow up email draft, or a chat you can push back on. Action items go straight onto a task board. The AI is instructed never to invent a decision, a task or a name that was not in the transcript.
06
It survives a crash
Meeting audio is written to disk continuously while recording, so a call is recovered on next launch if the app or the machine dies. If your microphone was muted or system audio was silent, it tells you within seconds, not an hour later.
MeetingsAccuracy Private Calendar
Where people use it
Or never leaves the room
On-device mode transcribes entirely on your own machine when you stop. Nothing is uploaded.
It knows what is next
Connect Google Calendar read-only, and the app spots when a call starts and offers to record. It never writes to your calendar.
Questions
Questions
Does a bot join my meeting?
No. Mr.Speech records on your computer, so other participants see no extra attendee. That also means it works with any conferencing tool, including ones that block bots.
Can I keep meeting audio entirely off the cloud?
Yes. Choose the on-device meeting mode and both the recording and the transcription stay on your machine.
Does it score participants or read the mood of the room?
No. There is no sentiment scoring, no engagement metric and no talk-time ranking presented as attention. We explain the reasoning on our EU AI Act page.
