Home / Technology / Google's Gemini Transcribe Masters 85 Languages
Google's Gemini Transcribe Masters 85 Languages
26 Aug
Summary
- Gemini 3.5 Transcribe boasts over 85 language auto-detection.
- The AI model formats unstructured speech and removes filler words.
- Future Chrome integration allows voice input in any web field.
Google has unveiled Gemini 3.5 Transcribe, an advanced AI audio model featuring improved speech recognition and transcription. This new model demonstrates superior performance over previous versions, with the capability to automatically detect more than 85 languages.
Gemini 3.5 Transcribe excels at transforming unstructured speech into formatted text, offering precise edits via voice commands and eliminating filler words. It also learns custom vocabulary and captures alphanumeric strings effectively.
This powerful AI can attribute speech to up to three speakers, providing word-level timestamps, which is ideal for transcribing content like podcasts. It powers features on Android devices and the Gemini app on macOS.
Looking ahead, users will soon be able to utilize this technology for voice input in any Chrome web field, enabling voice-dictated replies and prompts. Gemini 3.5 Transcribe is also being integrated into Google Antigravity, Search Live, Gemini Live, Docs, Keep, and Gmail, with API access available for developers.
