Google’s new AI app turns speech into text in real time—even offline: Know what it is

Google has released an AI-powered speech-to-text app called Google AI Edge Eloquent that can run offline on iOS devices. The app leverages the Gemma AI model to accurately recognise speech and convert it into text. As users speak, it will show a live transcription, and it also filters pauses for accurate text.

Alongside real-time transcription, the Google AI Edge Eloquent app can also transform the text into key points, formal, Short, and long. Here’s everything you need to know about the new Google dictation app.

What is Google AI Edge Eloquent?

Google AI Edge Eloquent is a dedicated app that converts speech into text in real time. Google’s app store listing says that Eloquent is an “advanced dictation app engineered to bridge the gap between natural speech and professional, ready-to-use text.” With the help of AI, Eloquent cleans the speech, cutting out fillers and corrections to deliver polished, accurate text.

How does Google AI Edge Eloquent work?

Once users download the Google AI Edge Eloquent app, they need to install the Gemma-based automatic speech recognition (ASR) models, and users can directly dive in to transcribe speech. Users can simply turn off cloud mode, which means the transcription happens locally on your device. In cloud mode, the app uses Gemini to clean up and improve text.

The AI model also adapts and learns user vocabulary by importing keywords, names, and jargon from your Gmail. In addition, users can also add their own custom words. Users will also have control to turn off cloud processing, allowing the app to run entirely on-device for greater privacy and offline use.

The app also showcases the user’s transcription history and lets users search and track metrics like speed and total words spoken. The Google AI Edge Eloquent now comes as a competitor to other transcription apps like Wispr Flow, SuperWhisper, Willow, and others.

Google AI Edge Eloquent availability

The Google AI Edge Eloquent app is currently compatible with iOS devices, with an Android version likely to launch soon. The App Store description further states that it offers “seamless integration,” so that it works across systems without needing to switch apps. It also comes with a floating button similar to Wispr Flow, allowing users to quickly start transcription from anywhere on their screen.