Transcribe Zoom meetings on Mac without bots

Zoom remains the standard video conferencing platform for millions of businesses, consultants, legal professionals, and software engineers. Yet recording a Zoom call traditionally forces professionals to make uncomfortable trade-offs. You either invite an awkward virtual attendee bot into the meeting room—which interrupts participants and triggers client compliance concerns—or you rely on Zoom’s cloud recording infrastructure, which stores proprietary transcripts on remote servers and requires expensive ongoing SaaS seat subscriptions.

Oyma eliminates these compromises. By combining native macOS ScreenCaptureKit audio capture with on-device OpenAI Whisper speech recognition and an open Markdown knowledge base, Oyma allows you to capture, transcribe, and synthesize Zoom calls silently on your Mac with zero cloud exposure.

The problem with conventional Zoom meeting bots

Most AI meeting utilities approach Zoom capture by deploying automated cloud attendee bots. While technically simple for vendors to build, this approach creates immediate operational and privacy hurdles:

• Client friction: When a virtual bot joins a Zoom call, it appears as an artificial attendee in the participant roster. Clients, prospective hires, and enterprise partners frequently question who owns the bot and where their voice data is traveling.
• Enterprise bot blocking: Increasing numbers of enterprise organizations configure Zoom administrative policies to automatically kick, ban, or refuse entry to third-party recording bots.
• Remote audio custody: Cloud bots stream raw audio across the internet to remote servers, exposing confidential strategic planning, legal discussions, and proprietary product roadmaps to third-party data processing risks.
• Disconnected summaries: Once the meeting concludes, summaries remain stranded in a vendor’s web dashboard rather than living inside your personal knowledge library.

How Oyma captures Zoom calls on macOS

Oyma takes an architectural approach rooted in local Apple Silicon capabilities. Instead of joining the call remotely, Oyma operates entirely within your Mac’s operating system:

┌────────────────────────────────────────────────────────────────────────┐
│                     ZOOM CAPTURE PIPELINE ON MACOS                     │
├────────────────────────────────────────────────────────────────────────┤
│ 1. Zoom Desktop Client opens active audio session                      │
│ 2. CoreAudio detector recognizes Zoom process & displays banner        │
│ 3. User clicks [Record Meeting] (explicit confirmation)                │
│ 4. ScreenCaptureKit captures system audio + mic into local buffer       │
│ 5. On-device Whisper transcribes speech on Apple Silicon Neural Engine │
│ 6. Formatted note writes directly to vault's meetings/ folder (.md)    │
└────────────────────────────────────────────────────────────────────────┘

1. Automated Zoom meeting detection

Implemented in src/main/meeting/meeting-detector.ts, Oyma monitors macOS CoreAudio audio session notifications. The moment the Zoom macOS application initializes an active VoIP audio stream, Oyma identifies the process signature and presents an unobtrusive notification banner:

Zoom meeting detected.
[Record Meeting]   [Dismiss]

You retain total control: recording never begins automatically without your explicit confirmation.

2. Silent two-channel ScreenCaptureKit recording

When you click record, Oyma engages Apple’s native ScreenCaptureKit framework. It captures both channels simultaneously:

• System audio: What you hear from other call participants through your headphones or speakers.
• Microphone input: Your own voice speaking into your Mac’s internal or external microphone.

Because capture occurs directly at the operating system audio pipeline, no virtual bot ever enters the Zoom room. Other participants see only you, and zero network traffic is dispatched to external servers.

3. 100% on-device Whisper transcription

Once the meeting finishes, Oyma processes the captured audio using optimized OpenAI Whisper models running locally on Apple Silicon. The transcription pipeline supports 13 languages with automatic language detection. Not a single packet of audio data leaves your device.

4. Direct Markdown output in your local vault

Instead of locking your notes inside a closed database, Oyma formats the transcript and executive summary into an open, human-readable .md file, saving it directly into your vault:

---
title: "Q4 Strategic Enterprise Review"
date: 2026-10-08
platform: Zoom
attendees: ["Sarah Jenkins", "Michael Chang", "Alex Rivera"]
tags: [zoom, enterprise, strategy]
---

## Executive Summary
Discussion centered on Q4 contract expansion and security audit timelines. Client approved the deployment schedule pending updated architecture diagrams.

## Action Items
- [ ] Sarah: Provide updated network topology diagrams by Friday
- [ ] Alex: Review data custody requirements with compliance team
- [ ] Michael: Schedule technical onboarding session for next Tuesday

## Key Decisions
- Deployment rollout will proceed on Apple Silicon infrastructure
- All customer transcripts will be retained locally in Markdown vaults

Because files are stored as plain text, you can open and edit them immediately in Obsidian, Visual Studio Code, iA Writer, or standard terminal utilities.

Practical workflow: From Zoom call to interconnected knowledge

Integrating Oyma into your daily Zoom routine requires zero configuration friction:

1. Join your call: Open your scheduled Zoom link in the native Zoom client.
2. Accept the prompt: Click the Oyma desktop notification or press your global recording shortcut.
3. Take live notes: Use Oyma‘s live-preview Markdown editor to jot down impressions, tasks, or reference links while the call proceeds.
4. End and synthesize: When you leave the Zoom meeting, stop recording. Oyma runs on-device Whisper transcription, synthesizes key takeaways using local Ollama or your chosen API keys, and weaves your manual notes with the spoken dialogue into a unified record.
5. Link to your graph: Add bidirectional [[WikiLinks]] to connect the meeting note to client profiles, active project specs, and your daily notes.

Verifiable privacy: The Egress Log guarantee

To verify that your Zoom discussions remain confidential, Oyma provides a built-in Egress Log that audits every outgoing network call from the application. You can review the ledger at any time to confirm that audio data, transcripts, and personal thoughts stay strictly on your Mac.

Experience silent, private Zoom meeting capture without attendee bots or cloud lock-in today.