level: business
google announced rambler, a new ai voice dictation feature for its gboard keyboard app, at the android show event. rambler uses gemini-based multilingual models to transcribe speech, remove filler words, and handle midsentence corrections. it also supports code switching, letting users mix languages like english and hindi without losing context. google says the feature clearly indicates when it is active and does not store voice recordings, using audio only for transcription.
the feature will first roll out on samsung galaxy and google pixel phones this summer, with wider android availability later. gboard is the default keyboard for most android users, giving rambler a built-in audience of hundreds of millions. this distribution advantage puts pressure on standalone dictation apps like wispr flow and typeless, which have mostly focused on desktop and ios. google also released an offline dictation app, ai edge eloquent, on ios last month, but rambler marks its biggest push into android dictation.
privacy is a key part of google's pitch. ben greenwood, director of android core experiences, said the company uses both on-device and cloud processing and has invested heavily in safety and privacy. this message targets users comparing rambler with third-party apps that may handle data differently. for dictation startups, the challenge is now to offer better accuracy, deeper features, or stronger privacy to convince users to download a separate app when a capable tool is already built into the keyboard.
why it matters: built-in ai dictation on gboard could reduce demand for standalone dictation apps, forcing startups to differentiate on accuracy, features, or privacy.