Apple and Google Just Started Transcribing Calls for Free. Neither One Touches a Zoom Tab.

The Feature Just Became Free

For most of the time MinuteAI has existed, “on-device transcription and summarization” was a reason to pick a third-party app over what shipped in the box. That’s gotten less true this year, and it’s worth being honest about exactly how.

iOS 26 built call recording and transcription directly into the Phone app: record a one-to-one phone or FaceTime audio call, and both the recording and a full transcript save into Notes, broken down by speaker, with an Apple Intelligence–generated summary layered on top — all processed on-device. Google shipped something with the same shape earlier: Gemini Nano running inside the Pixel Recorder app since 2024, generating three-bullet summaries of any recorded conversation entirely on the phone, with no round trip to a server. Two different platform vendors, converging on the same architecture MinuteAI has been built around from the start: don’t send the audio anywhere.

Where Both of Them Stop

The interesting part isn’t that this happened — on-device was always going to become table stakes once the silicon caught up. It’s where each vendor drew the boundary.

Apple’s version is explicit about its own limits: call recording covers one-to-one Phone and FaceTime audio calls, requires an Apple Intelligence–capable device (iPhone 15 Pro or newer, iPhone 16 or newer for the automatic transcript), isn’t available in the EU, Saudi Arabia, or South Africa, and — this is the part that matters most — does not work for calls made through WhatsApp, Zoom, Teams, Google Meet, or any other third-party app. Google’s feature is scoped just as narrowly in a different direction: Recorder transcribes and summarizes whatever the phone’s microphone picks up, which covers a lecture, an in-person conversation, or a call — but it isn’t wired into a video-meeting app’s audio pipeline either, and full functionality is gated to Pixel 8 and 9 series hardware.

Neither company shipped a transcription layer for the thing a huge share of actual work meetings run through: a browser tab open to Google Meet or Microsoft Teams on a laptop. That’s not a bug they haven’t gotten to yet — it’s a structural consequence of what each platform actually controls. Apple owns the Phone and FaceTime call stack end to end, so it can hook transcription into it. Neither Apple nor Google owns what happens inside a Chrome tab running someone else’s video-conferencing web app, so neither can quietly instrument it the same way.

Where MinuteAI Already Sits

That gap is the exact seam MinuteAI’s Chrome Extension was built to sit in. It captures the audio already playing in your own Meet or Teams browser tab and streams it to a local server on your own Mac — not a bot joining the call, not a cloud upload, no dependency on the meeting platform exposing an API for it. The rest of the pipeline stays on the same side of the line the platform vendors just validated: transcription runs on-device through Whisper via CoreML, summarization runs through a local GGUF model via llama.cpp instead of a cloud LLM, and speaker diarization runs through a locally-hosted pyannote.audio runtime rather than an API call. Apple and Google just spent a product cycle proving that architecture is the right one to bet on. They haven’t yet built the part of it that covers a desktop video call.

This Gap Won’t Stay Empty

It would be wrong to read that gap as permanent. Google Meet has already started expanding its own transcription and summary features to cover meetings that happen on Zoom and Teams too, not just Meet — aimed squarely at displacing the same category of tool, Otter and Fireflies included, that MinuteAI’s bot-free design was built to avoid competing with directly. That version runs through Google Workspace rather than fully on-device, which is a different privacy posture than either Recorder’s local summaries or MinuteAI’s local pipeline, but it’s a signal that the “video meeting in a browser tab” surface isn’t going to sit uncontested for long. Legacy Dragon, PrivateAI, and the rest of the AityTech portfolio all bet on the same thing MinuteAI does: that keeping data on the device isn’t just a feature to add once a market matures, it’s the reason to exist before it does.

What Actually Changed

The honest framing isn’t “MinuteAI is safe because the platforms haven’t caught up.” It’s narrower than that: Apple and Google both just spent real engineering effort validating that on-device transcription and summarization is the correct architecture, in the exact places they control outright — the phone call and the voice memo. The place they don’t control outright — a video meeting running inside somebody else’s web app in a browser tab — is still open, and it’s the one MinuteAI was built for from day one.

---

MinuteAI is available on the App Store for macOS and iOS, or you can learn more at getminute.app. Questions about how it fits your workflow? Reach out at hello@aitytech.com.