Meeting Coach

For people who work in English but did not grow up in it

You have no idea how you sound in meetings.

Nobody corrects a client. Nobody corrects their CTO. The mistakes that matter are invisible from the inside — a preposition that lands oddly, an answer that took ninety seconds when fifteen would have been stronger, hedging that turned a recommendation into a guess.

No account, no card, and nothing is stored.

Prepositions

Real findings, from a real client call.

Reads transcripts from

ZoomMicrosoft TeamsGoogle MeetOtter.aiFireflies.aiRead AIGranolaFathomtl;dvGrain

Three steps, about a minute

1

Paste the transcript

From Zoom, Teams, Meet, Otter, Fireflies, Read or Granola. It is already sitting in a folder doing nothing.

2

Say which speaker is you

Everyone else is pseudonymised before anything leaves our server.

3

Read the report

Your numbers, your English, how you came across, and three things to do differently next time.

Why this and not the others

Bring your own transcript

No bot in your call, no extension, no recorder, and no awkward conversation with a client. It works on meetings that already happened, and on meetings somebody else recorded.

It knows what your job is

“Was that a good answer for a CTO on a client pitch” is a different question from “was that grammatical”. Your role, your first language, and what you wanted out of the meeting all change the report.

Language and delivery together

How your English read, and how you came across. Did you answer the question you were actually asked? Did the number arrive first or last? Every other tool does one or the other.

A profile that accumulates

Per-meeting feedback is a novelty. The value is the five mistakes you actually keep making, measured per thousand words so meeting length cannot flatter you, moving down over months.

Your transcripts contain client conversations. We know.

This is the first thing our audience asks, so it is the first thing we answer.

Everyone else is pseudonymised

Before the transcript leaves our server, other participants become Speaker B and Speaker C. Their words are context. They are never analysed and never scored.

Zero training

Model providers are used under no-training terms. We say it here rather than burying it on page four of a policy.

Short retention by default

The raw transcript is deleted after 30 days; your findings and profile survive it. Or have it erased the moment your report is ready.

Delete everything, and mean it

One button removes every transcript, report, finding and your whole mistake profile.

What we will not do

We will not tell you that you made a mistake you did not make. Every finding quotes your own words, checked against the transcript before you see it — if the quote does not match, the finding is thrown away rather than tidied up.

Spoken English is not written English, so contractions, fragments, starting a sentence with “and”, and correcting yourself mid-thought are never flagged. And where a transcript simply cannot tell us something — your pace, whether anyone interrupted you — the report says so instead of guessing.

Find out how you sounded in your last meeting

Paste a transcript you already have. You will get your numbers and your three clearest English fixes in about thirty seconds.

Try it free

No account. Nothing stored.