Google launches Rambler, an AI-powered dictation feature for Gboard that competes with standalone dictation apps

Google announced Rambler at its Android Show: I/O Edition 2026 event on Tuesday morning. The feature uses Gemini-based multilingual models and removes filler words like 'ums' and 'ahs'; it also understands mid-sentence corrections and supports code switching (moving between languages mid-sentence without losing context). Rambler will be limited to Samsung Galaxy and Google Pixel phones for an initial summer rollout, but will eventually reach other Android devices.

Rambler uses a combination of on-device and cloud-based processing, does not store voice recordings, and uses audio only to transcribe speech. Gboard will clearly indicate to users when the Rambler feature is in use. The feature works across all apps on the device.

Rambler puts Google in direct competition with dictation startups including Wispr Flow, Typeless, Willow, SuperWhisper, Monoglogue, and Handy. Because Gboard is the default keyboard for the vast majority of Android users worldwide, Rambler arrives pre-installed for hundreds of millions of people — a distribution advantage that standalone apps must overcome with significantly better accuracy, deeper features, or stronger privacy guarantees to justify a separate download.