Record the meeting.
Keep the meeting.
Hit record and get a note with a summary, the decisions, the action items and the full transcript. Transcribed on your own Mac by a local speech model — no bot joins your call, no account, and the audio never leaves your machine. A Pro feature, macOS only.
A meeting you can
actually go back to
Recording is the easy part. What matters is what is in your vault an hour later — a note in the same plain markdown as everything else you write.
Transcribed on your Mac
Whisper runs locally, on the GPU. The audio never leaves your machine — not to us, not to a transcription service, not anywhere.
No bot joins the call
Nothing shows up in the participant list and nobody has to approve anything. Ampersand records the way a voice recorder on the desk would.
In person or on a call
Record just your microphone for a conversation in the room, or a call — where Ampersand also captures the other side through the system.
A note, not a wall of text
You get a summary, the decisions and the action items up top, with the full transcript folded into a toggle underneath.
Action items as checkboxes
What people agreed to do comes back as real task-list items in your note — the same ones the rest of the editor understands.
Who said what, on calls
Your microphone and the far end are two separate tracks, so a call transcript is labeled without guessing. No voice fingerprinting involved.
Speaks your language
Whisper handles dozens of languages and the summary comes back in the one that was spoken. Set it explicitly if automatic detection gets it wrong.
The audio expires
By default the recording is deleted the moment the transcript exists. Keep it for 7 or 30 days instead, or indefinitely — and delete any of it by hand.
The audio never
leaves the Mac
- Whisper runs locally, on your GPU
- Encrypted at rest with this Mac’s own key
- Deleted after transcription by default
From “can everyone
hear me?” to a note
The recorder lives inside the note you are writing, so it is never a panel floating over the screen you are sharing.
Type /record, or use the menu bar
Drop a recorder into the note you are already taking, or start one from the menu bar or ⌥⌘R when there is no window in front of you.
Pick in person or a call
In-person mode only ever touches the microphone. A call additionally captures the other side’s audio, which is the one mode that asks for screen permission.
Take notes while it runs
The recorder is a small card inside your document, not a window floating over your screen share. Keep writing around it; the transcript lands below.
Stop, and watch it write
Transcription has been running the whole time, so the summary starts appearing within seconds — streamed into the note line by line as the model writes it.
macOS only, for now
Free vs Pro
Good to know
Does my audio get uploaded anywhere?
No. Speech is transcribed by a Whisper model that runs on your own Mac, downloaded once and kept locally. The audio is stored outside your vault so it never syncs, is encrypted at rest with a key belonging to that machine, and is deleted by default as soon as the transcript exists.
What about the AI summary — does that leave my Mac?
Only if you have pointed it at a cloud provider. Ampersand summarizes through whichever AI you configured, so if that is Claude or OpenAI, the transcript is sent there with your own API key. Settings names the provider in plain text whenever the one you have chosen is off-device. Pick Apple Intelligence or a local model through Ollama and nothing leaves the machine at any point.
Does it work with Zoom, Meet, Teams, FaceTime?
Yes — with all of them, and with anything else that makes noise. Ampersand captures the audio your Mac is playing rather than integrating with a specific service, so there is no plugin to install and no meeting link to paste. It also means it works for a conversation in a room, which no meeting bot can do.
Do I need to tell people I am recording?
That depends on where you and they are, and it is your call to make. Ampersand is a tool on your computer, like a voice recorder — you are the one operating it. macOS shows its own orange microphone indicator the entire time a recording is running.
Can it tell people apart?
On a call, yes — your microphone and the far end are separate audio tracks, so “you” and “them” is a fact rather than a guess. In a room everyone shares one microphone, and rather than guess badly Ampersand leaves the transcript unlabeled and says so in the note.
How much disk does it use?
The speech model is a one-time download — around 570 MB for the accurate one, or 140 MB for a faster, rougher option. Recordings themselves are small and, on the default setting, deleted as soon as they have been transcribed. Settings shows exactly what is stored and lets you delete any of it.
What happens if I lose connection mid-meeting?
Nothing. Recording and transcription are entirely local, so a dropped connection has no effect on either. Only the AI summary needs the network, and only if you chose a cloud provider for it.
Does it work on Windows and Linux?
Not yet — recording is macOS only. It relies on the microphone and system-audio capture that macOS provides, and there is no equivalent we can ship on the other platforms today. On Windows and Linux the feature is absent rather than broken: nothing appears in the menu bar, in Settings or in the editor, and you are never offered an upgrade for it. Everything else in Ampersand is identical on all three.
Is meeting recording free?
Recording is part of Ampersand Pro, the one-time upgrade with no subscription. Reading a meeting note you already recorded is never gated — it is an ordinary markdown file in your own folder, and it stays readable whatever happens to your license.
Stop taking notes
about the meeting.
Be in the conversation and let the transcript take care of itself. Download Ampersand and record your next one.
Free to start · Mac, Windows & Linux · No account required