The Oats blog
Short, specific pieces about recording law, where meeting data actually goes, and how local transcription works.
Last updated 21 July 2026
For your profession
AI meeting notes for lawyers
What ABA Opinion 512, Model Rule 1.6(c) and NYC Bar 2025-6 require — and where on-device processing fits.
AI session notes for therapists
HIPAA business associates, the psychotherapy-notes carve-out, and what the APA, ACA and NASW codes require.
AI meeting notes for consultants
Client NDAs, GDPR Article 28 sub-processors, and the security questionnaire that asks who else touches the data.
Recording law
All-party consent laws, state by state
The 12 US states that require everyone's consent — each with statute, penalties and the nuance that catches people out.
Do you need everyone's consent to record a meeting?
The short answer, the exceptions that trip people up, and why an invisible notetaker doesn't change it.
Where your data goes
Where each AI notetaker actually stores your data
Otter, Fireflies, Granola and Oats compared on the question a compliance review actually asks — policy vs. architecture.
How local Whisper transcription works on Apple Silicon
Metal, CoreML, voice activity detection and audio mixing — less about the model than about what you refuse to send it.
Comparisons
Granola alternative
Both skip the bot. The difference is whether your audio goes to a cloud — and whether you pay $14 a month for it.
Otter alternative
What you trade when transcription moves from a server to your own machine, and what you keep.
Fireflies alternative
“We don't train on your data” versus “we never receive it.”
Meeting notes without a bot
The four ways to do it, and how device-audio capture works on macOS.
Local transcription on a Mac
What running Whisper offline actually takes, and what it costs you in speed.