Last updated: 2026-07-27
Released in April 2026 by Google's AI Edge team, Eloquent is an iOS dictation app built on Gemma based on-device speech models. It rewrites spoken audio into clean, disfluency-free text without sending audio off the phone by default, unlike cloud-first dictation tools that transmit recordings to a server.
About Google AI Edge Eloquent
Google AI Edge Eloquent is a dictation app built by Google's AI Edge team that launched quietly on iOS in April 2026. It turns spoken thoughts into polished prose instead of a raw verbatim transcript, automatically cutting filler words like ums and uhs and cleaning up mid sentence corrections. The app runs on Gemma based speech models that Google ships as part of its on-device AI Edge platform, addressing a gap between dictation tools that transcribe speech word for word and cloud services that add latency and a privacy tradeoff. Eloquent processes audio entirely on the iPhone once its Gemma based automatic speech recognition model downloads on first launch, so transcription and cleanup work with no network connection afterward. A cloud mode is available for users who want deeper text rewrites powered by Gemini, but it stays off by default so audio never has to leave the device unless a user turns it on. The app also builds a personal context dictionary from a user's own vocabulary over time, improving recognition accuracy for names and jargon a generic model would misspell. The tool suits people who think out loud and want a usable first draft without editing: journalists filing quick notes, students transcribing lecture recordings, and professionals dictating emails or messages on the move. It also transcribes existing audio and video files pulled from the Photos or Files app, doubling as a batch cleanup tool for voice memos recorded earlier. It is not built for multi speaker use, since a two person conversation comes back as a single undifferentiated transcript rather than a labeled meeting transcript. Eloquent is free with no subscription and no stated usage cap, distributed under Google's AI Edge developer umbrella rather than as a numbered Google One tier. It ships only for iPhone, iPad, and Apple Vision Pro running iOS or iPadOS 16.0 or later; Google has referenced an Android release and system-wide keyboard integration in App Store copy but has not published a ship date. Google folded Eloquent into a broader on-device push in 2026 that also brought the multimodal Gemma 4 12B model to laptop apps in June 2026. The most common complaint in user reviews is the lack of a system-wide keyboard, which forces copying cleaned text into other apps instead of dictating directly where a user is typing.
Pricing
No cost to download or use; every feature, including cloud rewrite mode, works without payment as of July 2026. It sits under Google's AI Edge developer program rather than a numbered Google One plan, and Google has given no commitment that pricing stays this way once the app leaves testing.
Key Features
- On-device speech cleanup: Runs a Gemma based ASR model locally on iPhone to rewrite spoken audio into polished text, cutting filler words and false starts without sending audio to a server.
- Audio and video file transcription: Accepts existing recordings from the Photos or Files app and transcribes and cleans them up at the same speed as live dictation, turning it into a batch tool for old voice memos.
- Personal context dictionary: Learns a user's own vocabulary, including names and jargon, from repeated dictation sessions to improve recognition accuracy over time.
- Optional cloud rewrite mode: Lets users opt into Gemini powered cloud cleanup for heavier text rewrites, while keeping the default fully offline so audio stays on-device unless the user turns cloud mode on.
- Live waveform transcription view: Shows a real-time transcript and waveform while recording, with a pause control before committing to the final cleaned output.
Pros
- Unlimited use at zero cost. There is no subscription, trial limit, or paywall as of July 2026, unlike Wispr Flow's metered free tier.
- Genuinely offline. Core transcription and cleanup run on-device with no network call required, unlike cloud-first competitors such as Otter.ai.
- Personal vocabulary learning improves accuracy on names and jargon after repeated use, a feature most free dictation apps skip.
Cons
- iOS only as of July 2026. There is no Android or desktop app despite App Store copy referencing an upcoming Android release, so Windows and Android users have no path to the app.
- No system-wide keyboard integration, meaning text has to be copied and pasted into other apps rather than dictated directly where a user is typing.
- Cannot separate speakers, so a two-person conversation returns as a single undifferentiated transcript, ruling out meeting or interview use.
Data Handling
- Training-data policy
- On-device processing by default; audio does not leave the device unless the user opts into cloud mode, which uses Gemini for text cleanup.
Frequently Asked Questions
How much does Google AI Edge Eloquent cost in 2026?
Nothing. Every part of the app, from live dictation to the optional Gemini cloud rewrite mode, is available at no charge with no ads and no tiered plan. There is no paid upgrade to buy, which sets it apart from most dictation competitors that gate advanced cleanup behind a subscription.
Is Google AI Edge Eloquent free to use?
Yes, every feature is free today, including offline on-device transcription, filler-word cleanup, and audio or video file uploads from Photos or Files. The real limits are platform, not price: it only runs on iPhone, iPad, and Apple Vision Pro on iOS 16.0 or later, and it cannot separate multiple speakers in one recording.
What are the best alternatives to Google AI Edge Eloquent?
Wispr Flow is the closest match for users who want cloud based reformatting plus Mac, Windows, iPhone, and Android apps. Superwhisper appeals to Mac users who want fully local speech recognition with adjustable Whisper model sizes. MacWhisper is a common pick for batch transcribing pre-recorded files entirely offline.
How does Google AI Edge Eloquent compare to Wispr Flow in 2026?
Wispr Flow is a paid app with a generous free tier and native support for Mac, Windows, iPhone, and Android, while Eloquent is free without limits but restricted to Apple devices. Pick Wispr Flow for cross-platform reach and heavier AI rewriting, or pick Eloquent if zero cost and on-device-only audio matter more.
How do you get started with Google AI Edge Eloquent?
Download Google AI Edge Eloquent from the iOS App Store, open it, and grant microphone access; it downloads its Gemma based speech model on first launch so it can work offline afterward. Tap record, speak naturally, then tap stop to have Eloquent clean up the transcript and copy it to the clipboard. No account or sign-in is required.