Meet smarter. Move faster.

Pre-meeting briefs, an in-meeting Advisor and bilingual minutes after — one loop. Every conclusion links back to the exact words and timestamp.

No language pre-selectionLocal-first captureNever used for training

Most of your meetings are hosted by someone else. Whether you get minutes depends on the edition they bought.

That isn’t our sales pitch — it is what Google and Microsoft each wrote in their own official documentation.

Google Meet official help documentation

The meeting organizer must hold a Workspace edition that includes “Take notes for me.” Whether the meeting produces minutes at all depends on which edition the organizer bought — not on you.

The same document, on languages

The feature supports one language at a time; multiple languages spoken in the same meeting aren’t supported. And of the 8 languages it does support, Chinese isn’t one.

Microsoft Teams admin documentation

The organizer can switch the in-meeting AI assistant off in one move — and that same switch takes recording and transcription down with it.

MLOG captures from your own machine: it doesn’t join and sends no bot, so it works whatever the other side uses. You still tell everyone you are recording — MLOG gives you the wording and writes that notice into the meeting record. Your minutes don’t depend on what the other side bought.

The four kinds of meeting MLOG gets used for first

They are not MLOG’s boundary — they are its starting point. The one thing they share: what gets said in the meeting is worth more than the meeting.

Cross-border sourcing negotiation

Sourcing lead · quarterly price talks with overseas suppliers
Today

Price, payment terms and MOQ get rewritten four times in an hour, in two languages at once. Afterwards you reconstruct it from memory and scattered notes; by the next round nobody can say for sure whether that number was locked.

With MLOG

Ask quietly when you’re put on the spot and the Advisor answers: “Last locked at FOB $12.40/unit, valid through Q4.” Transcription needs no upfront declaration of which language this meeting will be in; the moment it ends you get a team-language minute and an English recap — two separate documents, neither collapsed into the other’s dominant language.

Next to every number sits a timecode 6-30 · 00:17:26 — open it inside MLOG and you’re at the exact words and audio. What isn’t in the knowledge base gets a plain “not found.” It doesn’t guess.

Export sales follow-up

Sales rep · dozens of client calls a quarter
Today

Lead times promised, prices quoted, who owns the follow-up — scattered across dozens of notes and chat threads. To answer “what did we actually agree to,” you page back through all of them.

With MLOG

Ask once and get all of it: “What lead times have we promised this client?” Every answer carries the meeting it came from and the timestamp. Action items come out with an owner and a due date already on them.

The follow-up email stops at draft. It goes out when you press confirm — there is no automatic-send path in the system.

Remote work across time zones

Distributed team · the other side picks the time
Today

The call lands in the middle of your night, on their meeting software, their account, their policy. The colleague who couldn’t make it gets a minute the next day with no context and no way to ask a follow-up question.

With MLOG

Capture happens on your own machine: it doesn’t join, it sends no bot, and it doesn’t care which meeting software they use. You still announce the recording — MLOG hands you the wording in the local language and logs that you gave it. Whoever missed the call gets bilingual minutes plus the exact words behind every line.

Audio is encrypted to this machine before any processing; recording carries on through a dropped network, and a crash costs you ≤5 seconds. A meeting can’t be replayed — a failed capture is an incident.

Customer success review

Account owner · quarterly reviews and renewal prep
Today

One review means digging through three months of meeting records for “what exactly did they complain about back then.” What you find is somebody’s written conclusion. What you don’t find is what was actually said.

With MLOG

Five minutes before you walk in, this account’s whole history is compressed onto one page: settled, unsettled, promised. Editable, exportable.

Every compiled line links back to that one meeting, at that one second. We don’t do sentiment detection and we don’t score the mood of a meeting — that isn’t evidence.

From “we had a meeting” to “it got done”

01 — Before

Walk in with the right questions

Five minutes before the call, MLOG compiles this counterparty’s full history from your knowledge base — the price you locked, the terms still open, the numbers they promised — into one editable walk-in brief, each line linked to last meeting’s exact words.

  • Past decisions, open terms and key numbers, auto-compiled
  • Every brief item links to last meeting’s exact words and time
  • Editable agenda draft, exports to Notion / Google Docs
  • Syncs with Google Calendar / Outlook, ready before you join
Thu 10:00 · Acme renewal · Walk-in brief
Unit price locked last time at FOB $12.40 6-30 · 00:17:26
Your exact words: “No price change before Q4.”
45-day payment terms still open 6-30 · 00:52:18
Their finance insists on 60 days; your floor is 30
Worth asking: is the Q3 forecast confirmed 6-30 · 00:24:40
You promised the forecast “by Friday” — no record shows it was sent
Compiled from 3 past meetings & 12 docs · every line linked
02 — During

Answers when needed, never in the way

A quiet sidebar floating beside your call — no bots, nothing joins the meeting. It transcribes code-switching speech live. When the other side uses a term that is already in your glossary or your uploaded documents, it is added to the sidebar silently, with the source attached and at no cost. Ask when you’re stuck and it returns one answer, one data point and an evidence link. If it can’t find it, it says “not found.” It never pops, blinks or auto-expands.

  • Live code-switching transcription, searchable as you speak
  • One answer + one data point + an evidence link
  • Says “not found” when it can’t find it — never invents
  • Your credit balance is visible throughout; as it runs low only the Advisor steps back — capture and transcription carry on, recording never stops
In-meeting Advisor · never interrupts
$12.40/unit (FOB), locked through Q4
You asked: what did we quote them last time? — that is below their $13.10 ask by 6-30 · 00:17:26
No matching record found
You asked: when does their new line go live? — not in the knowledge base, and the Advisor does not guess
Credits this month0:24 / 3:00
03 — After

Decisions become bilingual minutes, actions and emails

The moment the call ends: a team-language minute, a client-language follow-up email, and action items with owners and due dates — generated together. Every decision carries a clickable timecode to the words and audio. Exports and sync stay drafts until you confirm; each meeting then settles into your knowledge base.

  • Bilingual minutes: team + client language in one pass
  • Action items with owners and due dates — nothing leaves until you confirm
  • Cross-meeting Q&A + knowledge base: ask “what did we promise?”
  • Exports & write-backs stay drafts — Notion / Google Docs await your confirm
Acme renewal · After-meeting outputs
FOB unit price locked at $12.40 through Q4 00:17:26
Team-language minute and English recap generated as a pair
MOQ reduced from 5,000 to 3,000 00:24:03
Action: send the Q3 forecast by Friday · owner ZQ 00:41:07
Follow-up email drafted · sends only after you confirmDraft

Capture on desktop, carry on mobile

One loop across every device you already use: desktop for online-meeting capture and the full workbench, mobile for in-person capture and the whole flow on the go.

macOS

Online-meeting capture and the full workbench. System-level dual-channel audio — no bots, nothing joins your call.

Download
Windows

The same full workbench as macOS, covering Zoom and other online meetings — local-first capture all the same.

Download
iOS

In-person capture on the go: brief, live Q&A and post-meeting confirmation — the whole flow on your phone.

Not on the App Store yet. The desktop app captures the same meetings, and the web app works in any browser.

Android

The same full flow as iOS — never miss a site visit or a trade-show meeting.

Available on Google Play.

Get it on Google Play

Powerful, and accountable

Meeting records carry trust. Every design choice in MLOG answers one question: would you act on this decision?

Local-first

Audio is encrypted to disk before processing — losing ≤5s on a crash. Meetings can’t be replayed; a failed capture is an incident.

Evidence before conclusions

Every AI conclusion carries a timecode link; if it can’t find it, it says “not found” — never pretty but unaccountable output.

Human in the loop

Outbound sends, exports and write-backs always need your confirmation; the system never sends on its own. No sentiment scoring.

Never used for training

Your audio, transcripts and minutes never enter any model training; vendors are bound by zero-retention terms too.

Start with your next meeting

Create an account, install the desktop app, and the loop closes itself — a brief before, the Advisor during, bilingual minutes after. Every conclusion links back to the words and the timestamp that produced it.

Creating an account is free and carries no allowance by itself. You pick the plan and click the payment yourself, and you can cancel at any time.