Google has released a new dictation app called Google AI Edge Eloquent on iOS, offering offline-first speech-to-text capabilities as it expands into a growing category of AI transcription tools.
Offline First Dictation Powered By On Device Models
The app uses Gemma-based automatic speech recognition models that can run locally once downloaded, allowing users to dictate without requiring a constant internet connection.
Users can view live transcriptions as they speak, and when recording pauses, the app automatically removes filler words such as “um” and “ah” while refining the text into a more polished format.
The app also offers transformation options such as “Key points,” “Formal,” “Short,” and “Long,” enabling users to adjust the output style of their transcriptions.
Cloud Mode Adds Additional Processing Options
Eloquent includes an optional cloud mode that uses Gemini models for additional text cleanup. Users can toggle this feature off to keep all processing local.
The app can also import keywords, names, and jargon from a user’s Gmail account, and allows manual addition of custom vocabulary.
Session Tracking And Performance Metrics Included
The app stores transcription history and allows users to search past sessions. It also provides metrics such as words spoken, speaking speed, and session summaries.
These features are designed to give users insight into their dictation patterns and improve productivity.
Android Integration Planned Despite iOS Only Launch
Although the app is currently available only on iOS, its App Store listing references future Android support. Planned features include the ability to set the app as a system-wide keyboard and access dictation through a floating button interface.
Google has removed earlier references to Android from its listing but noted that an iOS keyboard version is expected to launch soon.
Competition Grows In AI Transcription Space
Eloquent enters a market that includes apps such as Wispr Flow, SuperWhisper, and Willow.
The release reflects increasing demand for AI-powered transcription tools as speech recognition models improve in accuracy and usability.
Featured image credits: Flickr
For more stories like it, click the +Follow button at the top of this page to follow us.