SpeakON Launches a MagSafe-Compatible AI Voice Button Featuring a Built-In Microphone: Transforming Your Speech into Refined Communication and Cross-App Actions

Voice dictation on smartphones has been a solved problem for years — but the output has not been. Most tools return a verbatim transcript, fillers and false starts included, which the user then has to clean up and move somewhere useful. SpeakON, a new hardware product launching in the United States at $129, targets that gap directly: a 25 g magnetic button that snaps to the back of an iPhone, records with its own microphone, and delivers finished, shaped text straight into whatever app is already open. The company frames the category as an "AI Communicator" — inspired by the Communicator in Star Trek — aimed at founders, managers, consultants and field professionals who have plenty of ideas but no time to stop, unlock a phone, switch apps and type.

The Hardware: Own Mic, Own Battery, Own Storage

The device measures 58 x 58 x 6 mm and weighs 25 g. A 220 mAh battery is rated for 10 or more hours of continuous use and more than 2 weeks of standby, recharging in under 1.5 hours over USB-C. It carries 128 MB of onboard storage, a capture range of roughly 60 cm, and supports up to 5 minutes of continuous input per press. The box includes a USB-C cable and a magnetic ring for iPhones without MagSafe; magnetic attachment requires iPhone 12 or newer.

The defining architectural choice is that audio is captured by the button, not the iPhone's system microphone. That has three consequences: the phone mic stays free for CarPlay, calls or FaceTime; the app never needs continuous background microphone access, one of the permissions users most dislike granting; and because the button has its own battery and storage, it keeps working when the phone is locked or offline, buffering on-device and syncing when a connection returns.

How the Text Lands in Your Apps

SpeakON installs as a system-wide iOS keyboard extension (requiring iOS 16.0 or above). That mechanism lets output appear directly inside Messages, Mail, Slack, Notion or any text field, with no app switching or clipboard round trip — press, speak, and the processed text appears in the active field, ready for review before sending.

The software is a text-shaping layer rather than a verbatim transcription tool. Smart Polish strips fillers, restarts and redundant phrasing. Smart List detects sequence intent and returns structured bullets or to-dos. Style detects the destination app and adapts register, so the same spoken sentence lands casual in Messages and professional in Mail. Translation delivers text in 12 languages directly into the active field. Dictionary persists names, jargon and preferred spellings across sessions, addressing the proper-noun problem that plagues most voice tools. Voice Edits revises existing output by speech, and Notes stores captures as titled, searchable entries — including offline captures. All features are available now, with no subscription.

Privacy and Pricing

SpeakON states that voice data is encrypted, never sold and not used to train AI models, with user control over cloud sync. The company reports SOC 2 Type II, HIPAA and GDPR compliance.

The device sells for $129 one time in the United States, with Pro Lifetime included and no recurring fee. Colorways are Pearl, Black, Navy Blue, Amber and Sage. Shipping covers the 48 contiguous states plus Washington, D.C., with a 30-day return window and a 12-month warranty. The companion app is a free download and works without the hardware — the cheapest way to test the text-shaping layer before buying the button.

What Comes Next

The current stage is a text-output product. The next step, SpeakON Agent, arrives in October 2026 and extends the same press gesture from producing text to preparing reviewable Notes, Tasks and user-confirmed Actions — a move from dictation toward delegation. The question to watch is whether shaped, app-aware voice output can carve out real daily use where plain transcription never did.