Question 1

Does it work with Zoom, Google Meet, and Teams?

Accepted Answer

Yes. NotchLive captures system audio via ScreenCaptureKit, so it works with any app that plays audio - Zoom, Google Meet, Microsoft Teams, YouTube, Spotify, and more.

Question 2

Does it need an internet connection?

Accepted Answer

Only for the initial AI model and translation language pack downloads. After that, everything runs on-device with no internet required.

Question 3

What languages does it support?

Accepted Answer

Whisper AI transcribes speech in 90+ languages with automatic detection. For best results with non-English audio, use the Medium or larger model. Translation to 20 languages is available via Apple Translation (requires macOS 15+).

Question 4

Does it work on Macs without a notch?

Accepted Answer

Works great! The caption bar appears at the top center of your screen on any Mac.

Question 5

Is it on the Mac App Store?

Accepted Answer

NotchLive is available as a direct download only. ScreenCaptureKit system audio capture requires distribution outside the Mac App Store.

Question 6

Does it require Apple Silicon?

Accepted Answer

Apple Silicon is recommended for best performance with on-device AI, but Intel Macs are also supported.

Question 7

How is NotchLive different from Apple Live Captions?

Accepted Answer

NotchLive can display captions in the MacBook notch or in a floating Focus or Timeline window. Timeline lets you scroll, search, copy, and export the current live transcript while captions continue arriving. NotchLive also adds translated speech, session recording, and saved transcript export. Both use on-device AI for privacy.

Question 8

Does NotchLive identify people by their voice?

Accepted Answer

No. NotchLive assigns anonymous, meeting-local speaker labels. It does not create persistent voiceprints or claim to know someone’s identity from their voice.

Question 9

Can I correct speaker labels?

Accepted Answer

Yes. Participant Review lets you rename speakers, merge duplicates, reassign speech, split an utterance, or leave uncertain speech unresolved.

Question 10

Does speaker attribution run locally?

Accepted Answer

Yes. On supported Apple silicon Macs, remote-speaker attribution runs locally alongside NotchLive’s on-device speech recognition.

Question 11

Is Meeting Analytics generated by AI?

Accepted Answer

No. Meeting Analytics is calculated directly from the transcript and reviewed participant map. It does not ask a language model to invent a score.

Question 12

What is the difference between Meeting Analytics and Meeting Intelligence?

Accepted Answer

Meeting Analytics measures participation and attribution from the meeting record. Meeting Intelligence uses supported on-device Apple Intelligence features to generate topics, decisions, actions, entities, and summaries.

Question 13

What can Codex and Claude access through NotchLive?

Accepted Answer

Only the data surfaces you explicitly enable through NotchLive’s read-only local MCP server. If active Timeline access is enabled, Codex or Claude can follow finalized captions and translate the same stream into multiple requested languages. You can disable the server or individual data surfaces at any time.

Question 14

Does version 1.9 work on Intel Macs?

Accepted Answer

Core live captions continue to work on supported Intel Macs. Automatic speaker attribution requires Apple silicon, and features that depend on that attribution require its reviewed result.