Record every meeting.
Transcribe locally. Summarize instantly.
Verbatim is the native macOS meeting assistant that automatically detects your calls (Google Meet, Zoom, Teams, Webex), captures dual-channel audio with zero meeting bots, transcribes speech offline, and generates structured Executive Recaps with 100% on-device AI. Zero subscriptions. Total privacy.
Product Architecture Sync
- Zero external drivers: Eliminated virtual audio cables and kext dependencies using native macOS APIs.
- Echo Guard™ & Voice Enhancer: Automatic speaker acoustic isolation and studio-grade vocal mastering (5-band EQ, noise gate, brickwall limiting).
- 100% Offline Processing: Speech-to-text and executive summaries run directly on the Mac processor with zero cloud queries or API fees.
- Finalize standalone signed package for macOS Sonoma and Sequoia.
- Maintain one-click export with automatic Finder reveal.
Built for those who value clarity and privacy.
Say goodbye to fragile audio routing setups and costly monthly subscriptions. Verbatim combines Apple's native frameworks with compact, highly-optimized on-device AI.
Autonomous Meeting Detection • Zero Meeting Bots
Automatically detects active video calls across Zoom (native app & browser), Google Meet (Chrome, Safari, Arc, Edge, Brave), Microsoft Teams, and Cisco Webex. Prompts you with an instant native banner with the real meeting topic pre-filled. Never sends awkward virtual AI bots to crash your client meetings or request host recording permissions.
Echo Guard™
Listening to meetings through your Mac's built-in speakers? Echo Guard acoustically isolates the microphone capture signal to prevent audio feedback loops and speaker bleed.
Dual-Channel Recording via ScreenCaptureKit
Simultaneously record your voice from any microphone (built-in, AirPods, or external USB) and your Mac's internal system audio (Zoom, Microsoft Teams, Google Meet, videos, or music). Powered directly by Apple's ScreenCaptureKit: no kernel extensions (kexts), no unstable virtual audio loopback drivers like BlackHole or Soundflower.
AI Voice Enhancer
One click to turn raw audio into studio-quality speech. Applies neural background noise removal, 5-band vocal EQ, dynamic broadcast compression, and a -1 dBTP brickwall peak limiter.
Multi-Recording Merge & Waveform Trimmer
Had a call cut short or need to slice out silence? Enter Merge Mode to fuse up to 5 sessions chronologically, or drag graphical waveform cursor handles to crop any recording with live audio preview and automatic transcript slicing.
On-Device Searchable Speech-to-Text
High-precision neural speech recognition running directly on Apple Silicon and Intel hardware. Features automatic multi-language detection, time-aligned segments, and instant full-text search across all your archived recordings.
Executive Recap with Local AI
When recording finishes, an embedded on-device language model analyzes the transcript and distills it into an executive-ready brief: Executive Summary, Key Takeaways, and Decisions & Next Steps. Exported cleanly in Markdown ready for Notion, Obsidian, Slack, or email.
Direct Finder Export
Native audio encoding into 192 kbps stereo MP3. One click exports the recording to your Downloads folder and automatically highlights the file in macOS Finder.
System Requirements & Architecture
Built with modern SwiftUI, AVFoundation, and low-level C audio compilers to maximize efficiency on Apple hardware.
Everything you need to know about Verbatim.
Traditional meeting tools (such as Otter, Read.ai, or Fathom) require an external bot to join your Google Meet, Zoom, or Teams call as an extra attendee, which can feel intrusive, distracting, and frequently triggers privacy concerns from clients or requires host recording permission. Verbatim is 100% bot-free and invisible: it captures audio locally using Apple's native ScreenCaptureKit framework directly on your Mac, so no bots ever join your meeting and your audio never leaves your machine.
Verbatim actively monitors your Mac for incoming or active calls across native Zoom (including Zoom Workplace and pre-call screens), Microsoft Teams, Cisco Webex, and browser-based Google Meet sessions in Safari, Google Chrome, Arc, Brave, and Microsoft Edge. The moment a call starts, Verbatim displays a clean native notification banner with the real meeting topic pre-filled, letting you start dual-channel recording with a single click.
Yes! With Verbatim's Multi-Recording Merge feature, right-click any existing recording and select "Merge with other recordings". Check off up to five other sessions. Verbatim automatically stitches the audio files together in strict chronological order based on their original start timestamps, and generates a completely fresh, unified speech transcript and local Executive Recap for the consolidated master session.
No, absolutely not. Verbatim utilizes Apple's modern ScreenCaptureKit APIs introduced in macOS Sonoma. This captures system audio cleanly and securely from active applications (Zoom, Teams, Meet, browsers) without virtual audio cables, system reboots, or root privileges.
Never. Verbatim's entire architecture is engineered for 100% on-device execution. Both speech recognition and summary generation run directly on your Mac processor. No audio bytes, metadata, or text snippets ever leave your machine.
Verbatim includes dynamic hotplug detection for all connected audio hardware. Whenever you connect or disconnect AirPods, Bluetooth headsets, or USB studio microphones, the app automatically updates the input selector and recalibrates levels in real time.
The Executive Recap is summarized locally within seconds after recording ends using an embedded on-device language model. Output is organized into clear sections (Executive Summary, Key Takeaways, Decisions & Next Steps) and can be copied to your clipboard in clean Markdown with one click.
A Mac running macOS 14.0 (Sonoma) or macOS 15.0 (Sequoia) or later. The application is universally compiled for peak performance on both Apple Silicon (M1, M2, M3, M4) and 64-bit Intel Mac computers.