Features
Sextant records your calls on your Mac and turns them into a transcript, highlights and a summary you can check against the recording.
Recording
Sextant records the call from your Mac: your voice, the other side and what is on screen, and keeps you informed while it does.
It asks first
When a call starts in Zoom, Microsoft Teams, Slack, Discord, FaceTime, WhatsApp, Telegram, Webex or a browser, a small window asks whether to record. Nothing is captured until you click Record. Each app can be set to Ask, Always or Never.
Two tracks and the picture
Your microphone and the call audio are recorded separately, so what you said is always yours. The video is the call window and, while you present, the screen you share; never another app.
Health you can see
While recording, indicators show that the microphone, the call audio and the video are coming in. If one stops, you are told during the call, and the call page lists what is missing afterwards.
Safe against crashes and device changes
Files are written in small pieces as the call goes, so a crash costs at most the last few seconds. Recording follows your audio devices when headphones connect or disconnect, and keeps sound and picture in sync.
Tell the call
One click copies a short message saying you are recording, to paste in the call chat, and marks the call as announced. Sextant never types into the call app itself.
In person and imported
Record a meeting in a room with the microphone only, every voice told apart. Or drop an audio or video file on the window to get the same transcript, speakers and highlights.
Transcript and speakers
A transcript you can trust enough to quote, with the right names on it.
Live, then refined
The transcript appears while the call runs. After the call a second, more careful pass on your Mac improves the text and the speakers.
Every word is a link
Click any word to play the call from there. The word being spoken is underlined as it plays.
Names that stick
Name a voice once and Sextant recognises it in later calls as “Probably Sarah Chen”, with a match score, until you confirm. Wrong guesses are fixed in two clicks, with Undo.
Your vocabulary
Add product names, people and acronyms so the recognizer spells them right. Nothing is corrected afterwards by hand-written rules; the recognizer decides with the audio.
Editable
Fix a word, split or merge a turn, or move it to another speaker.
English
Sextant transcribes English. Other languages are not supported yet.
Highlights
The moments you would otherwise have to find again: decisions, commitments, open questions, dates.
Found for you
After the call, a language model on your Mac reads the transcript and marks decisions, action items, questions and deadlines.
Seen on screen
When someone starts sharing or the slide changes, Sextant notes it from the video.
Marked by you
Press a shortcut (⌃⌥H by default) or the menu bar button during the call, with an optional note.
With the picture
Every highlight keeps a picture of the screen at that second, the quote and the time. Click it to play the moment.
Across all calls
The Highlights page gathers them from every call, by day, filtered by kind, person or call.
Summaries you can check
A summary is only useful if you can trust it. Sextant shows where every line came from.
Cited, line by line
Each line ends with the times it came from. Click one to play the moment and see the quote, who said it and the screen at that second.
Checked against the transcript
Sextant compares every line with the moments it cites. A line it cannot support is marked “not verified in transcript”, so you know which ones to read twice.
Templates
General, 1:1, Standup, Sales or customer call and Interview are built in. Write your own, and choose a template per app.
Your notes count
Notes typed during the call shape the summary written after it, in a section of their own.
Search
Find what was said, even when you do not remember the exact words.
Words and meaning
Search matches the words you type and also passages with the same meaning, across every call. It runs on your Mac.
Straight to the second
Results are grouped by call, show who spoke, and open the call at that moment. ⌘K searches from anywhere in the app.
Library, storage and export
Recordings with video take space, and notes belong where you keep your other notes. Both are handled.
A library you can scan
Calls are listed by day with a one-line summary, the app, the people and the number of highlights. Filter by app, person, date or highlights.
Storage under control
See how much each call uses. Cleanup rules remove video or audio after the number of days you choose, per app if you like; transcripts, summaries, notes and highlight pictures are always kept. Nothing is removed without a preview.
Export
Markdown, JSON or SRT subtitles for any call. Automatic export writes a Markdown note for every finished call into a folder you choose, and leaves notes you have edited there alone.
Plain files
Each call is a folder with its audio, video, transcript, summary and highlight pictures, in your Library folder.
Works around your battery
With “Process calls only when plugged in”, the work after a call waits for the charger. Recording itself is never delayed.
Third-party transcription, if you want it
Off by default. For calls where you want a second transcript from a cloud service, Sextant can send the audio to ElevenLabs with your own API key.
Your key, your account
The API key stays in the macOS Keychain. ElevenLabs bills you directly for the audio it transcribes; Settings shows how much audio was sent each month.
Only audio
Your microphone and the call audio are uploaded after the call, with your vocabulary terms if you have any. Never video, screenshots, notes, names or summaries.
Nothing lost
The transcript made on your Mac is always kept. You choose which version a call shows, and can compare them side by side.
Current limits
What Sextant does not do yet, so you know before you rely on it.
- English only.
- Macs with Apple silicon only; Intel Macs are not supported.
- On very long calls (tested at 130 minutes) the written summary does not yet cover the end of the call. The transcript, speakers and highlights do.
- Speaker recognition has been measured on meeting-room recordings; its accuracy on call audio has not been measured yet.
- No calendar integration: call titles come from the call app, and names from your own naming.