A macOS menubar meeting notes app for confidential client calls

By Kyle Nelson, Founder, Fazit

If you're searching for a macOS menubar meeting notes app, pick one by audio path first. Fazit is built for confidential 1:1 client calls: it sits in the Mac menu bar, captures the call locally, transcribes on the Mac, and writes a finished Markdown note into your Obsidian vault without a bot joining the meeting.

That matters if your Tuesday is three client calls, a half-written follow-up email, and a confidentiality clause you actually have to live with. A meeting notes app can save the admin work after the call, or it can create a second record of the conversation that you now have to control.

macOS menubar meeting notes app for client calls

A menu-bar meeting notes app should be fast to reach, quiet during the call, and clear about what it stores. Fazit runs as a native macOS menubar app for people who run 1:1 client calls: consultants, coaches, lawyers, financial advisers and therapists.

According to Fazit's own architecture notes at getfazit.com, capture is bot-free. A Core Audio process tap reads the call app's output and the microphone directly. Nothing joins the meeting, and no other participant sees a new attendee.

The working-day version is simple. You're on a Zoom call with a client about a delayed project. You speak into your mic. The client speaks through Zoom. Fazit keeps those as two streams, "You" and "Them", then writes the note into your vault after the call.

Fazit is macOS-only. The product statement says there is no Windows, Linux, iOS or Android version, so this is a Mac answer to a Mac search query.

best macOS menubar app for meeting notes in Obsidian

If Obsidian is where your client knowledge lives, the note should arrive as a file. Fazit writes a plain Markdown file into a folder of your Obsidian vault, or into Apple Notes. There is no Fazit database holding another copy of the note.

That file-based workflow matters in ordinary practice. A coach can open last month's client note, see the commitments and due dates, and write the next session plan inside the same vault. A consultant can search across Markdown files with the tools they already use.

A finished Fazit note is designed around the client-call job: summary, notable moments, action items with owners, and a paste-ready follow-up email. The related Fazit post on client call notes in Markdown says a client-call note should capture the client, date, objective, decisions, commitments, owners, due dates and follow-up email.

The opinion here is blunt: Obsidian meeting notes should be files. Fazit's post on Obsidian meeting notes as files makes the same point: a meeting note is only yours if it is a file you can grep.

macOS menu bar meeting notes apps compared by job

There are several Mac menu-bar tools around meetings. The search result can look crowded because "meeting notes" can mean quick scratch notes, calendar join buttons, recording, transcription or finished client-session notes.

App or sourceWhat the source says it doesBest fit from the stated facts
FazitNative macOS menubar app that turns 1:1 client calls into Markdown notes in Obsidian, with local transcription and no audio fileConfidential 1:1 client calls where the note belongs in your vault
MenuBar NotezApple's App Store listing says it is a macOS menu-bar notes app for quick notes including meeting minutesManual quick notes from a menu-bar window
Fellow macOS appFellow's documentation says it provides menu-bar access to meeting notes, agendas, action items, video-call join links and task updatesCalendar-linked meeting workflow and team follow-up
MeetingBarMeetingBar's site says it displays calendar meetings and supports one-click joining for video meetingsGetting into the right meeting quickly
Notes4MeThe GitHub page describes a privacy-focused macOS menu-bar meeting notetaker that records, transcribes and summarizes meetings using local AI processingLocal recording, transcription and summaries
TalkFoldTalkFold's site says it is a native macOS meeting-notes and transcription app with on-device transcription and quick recording from the menu barOn-device transcription with menu-bar recording

The comparison point is narrow. Menu-bar placement tells you where the button lives. It doesn't tell you whether a bot joins, whether audio is recorded, whether transcription runs locally, or whether the final note is a Markdown file in your vault.

For a lawyer's client call, a financial adviser's review call, or a therapist's session note workflow, those storage questions matter more than the icon in the top-right corner of the screen.

meeting notes app for Mac without a bot

Fazit does not join the call. That means there is no meeting participant named after a note-taking vendor, no calendar bot to admit from the waiting room, and no third-party attendee visible to the client.

The capture path is local. Fazit taps the call app's audio and your microphone through Core Audio. The README for Fazit describes the 1:1 attribution trick: your microphone is "You", and the call app's output is "Them". Because the two streams are physically separate on the Mac, Fazit doesn't need multi-speaker diarization for a two-person call.

That shows up in the workday. If a consultant says, "I'll send the pricing model by Friday," and the client says, "I'll send the usage data tomorrow," the transcript can preserve who said what because the two sources were separate from the beginning.

Fazit deliberately targets 1:1 calls. The research notes say that with three or more participants, every remote voice is mixed into one system-audio channel, and separating them requires diarization. The product's clean attribution claim is for 1:1 client calls.

local transcription on Mac for confidential calls

Fazit transcribes on the Mac. The site's architecture statement says transcription runs on-device using Parakeet via CoreML, and note generation runs through a local model on localhost. Call audio and transcript content do not leave the machine for note generation.

The source also draws a line that is easy to miss. "Nothing leaves the Mac" is a claim about call content. Models are downloaded once from public CDNs on first run, and account, licensing and payment use Fazit's servers. Those are separate from the call-content path.

That distinction is the kind a security reviewer can actually use. A financial adviser can ask, "Does the conversation go to your servers?" Fazit's answer is about the content path: call audio, transcript and notes for generation stay on the device.

For the engineering background, Fazit's post on on-device transcription on Apple Silicon explains the larger shift: speech recognition moved to the cloud when laptops couldn't run the models, and Apple Silicon changed that constraint.

meeting notes app that does not write call audio to disk

Fazit's strongest architectural claim is narrower than "private". The call audio is never written to disk.

According to Fazit's architecture notes, audio exists only in a fixed-size RAM ring buffer with no write API. The buffer is destroyed on every exit path, including errors. Every note records audio_retained: false in its frontmatter.

That gives you a different artifact after the call. You have a Markdown note and, if enabled by the workflow, a transcript in the note. You don't have an audio file sitting in a vendor account, a downloads folder, or an app database.

A therapist finishing a client session should not have to remember whether a recording file was created and then deleted later. A lawyer should not have to ask whether a vendor still has audio in a processing bucket. Fazit's architecture is designed so the audio file is never created.

The related post on AI meeting notes without recording puts the test plainly: no bot, no cloud transcription, no audio file, and no vendor database holding the note.

macOS menubar meeting notes app and consent duties

Bot-free capture does not remove consent duties. Nothing here is legal advice, and consent and professional-conduct duties apply to the person running the call.

Fazit's own accuracy notes say the posts describe architecture and cite primary sources, while consent and professional-conduct duties still apply to the person running the call. The README says Fazit's architecture reduces the vendor's exposure and the audio-file exposure, and it does not remove the user's duty to have consent where law requires it.

That is the practical line to use before a client call. If your jurisdiction, contract, regulator or professional rules require notice or consent, you handle that before capture starts. The cleanest product architecture does not speak the sentence for you.

For a US consent-law overview, Fazit has a separate post on recording client calls and consent laws. For lawyers, the privilege-specific discussion is in AI scribes for lawyers.

what to check before choosing a macOS menubar meeting notes app

Use this checklist before you put client conversations into any meeting notes app:

  • Does a bot join the call as a participant?
  • Does the app require calendar access to detect meetings?
  • Does transcription run on the Mac or on a vendor server?
  • Is any call audio written to disk?
  • Is the final note a Markdown file in your vault?
  • Is there a vendor database holding another copy of the note?
  • Can you separate call content from account, licence, payment and update traffic?
  • Does the app state plainly that it makes no legal guarantee?
  • Does your own consent script match your jurisdiction, contract and professional duties?

A consultant under an NDA should start with the third-party disclosure question. A lawyer should start with privilege and client notice. A therapist should start with consent, professional duties and whether a session ever leaves the practice environment. The tool choice follows those answers.

Fazit's answer is specific: no bot, local transcription, localhost note generation, Markdown output into Obsidian or Apple Notes, no Fazit database copy, and call audio never written to disk.

Sources