Google on Monday quietly released an offline-first dictation application, titled “Google AI Edge Eloquent”, for iOS, entering a growing field alongside rival services.
According to TechCrunch, the application is available free of charge. Once its Gemma-based automatic speech recognition models are downloaded, users may begin dictation on their devices. The interface provides live transcription, while a pause function refines the text by removing filler expressions and improving clarity.
Options beneath the transcript allow users to adjust the output, including “Key points”, “Formal”, “Short”, and “Long” formats.
The application also permits local-only processing by disabling cloud mode. When enabled, cloud mode relies on Gemini-based models for further text refinement. Users may import selected keywords, names, and technical terms from their Gmail accounts, while additional custom entries may be added manually.
A record of transcription sessions is maintained within the application, enabling users to review and search previous entries. It also provides data such as words spoken during the last session, speaking speed, and total word count.
In its App Store description, the company states that the application is designed to convert natural speech into polished written text, removing hesitations and mid-sentence corrections to produce clear prose.
At present, the application is limited to iOS, although references to an Android version are included in the description. Further details have been sought from Google.
The description adds that Android integration would allow the application to function as a default keyboard, offering access across text fields. A floating button feature, similar to existing tools, is also anticipated to enable quick transcription access.
The release comes amid rising interest in AI-powered transcription tools, as speech recognition technology continues to improve.
.jpg)
