How to install Google announces Gemini 3.5 Transcribe for AI-powered

What happened

Google has announced Gemini 3.5 Transcribe, a new AI-powered speech-to-text model that extends the technology already running inside Gboard's Rambler feature to a broader set of Google products. The announcement positions Gemini 3.5 Transcribe as a serious option for developers and end users who need accurate, on-device or cloud-assisted transcription without reaching for a third-party service. Chrome is among the first non-Gboard destinations confirmed to receive the capability, signaling that Google intends this to become a foundational layer across its ecosystem rather than a one-off feature.

If you are already using Gboard's Rambler or are building speech-capable apps on Google's developer platform, this guide covers what arrived, what changed in the API surface, how to get the update, and where the rough edges are right now. Nothing here is speculative — the facts come directly from Google's announcement as covered by Ars Technica.

How it works

Gemini 3.5 Transcribe is the model powering AI-driven speech-to-text inside Google's products. It was previously the engine behind Rambler, the long-form dictation feature in Gboard. Google has now formally named and surfaced the model so that other product teams and external developers can call it by name and integrate it explicitly. Chrome is the first confirmed expansion target beyond Gboard.

The main change for developers is that the transcription capability now has an explicit, addressable identity as Gemini 3.5 Transcribe rather than being an opaque system service. That matters because API clients, SDK integrations, and app configurations can now target this model by name, which allows builders to reason about which model is handling transcription, swap between models if Google releases alternatives, and write documentation that is specific rather than vague. If you were previously relying on the general speech recognition service inside Google's mobile or browser infrastructure, you should expect the path forward to route through the Gemini model family instead. Builders working with Chrome will gain access to transcription through whatever surface Google exposes inside the browser's web speech or AI APIs as the rollout proceeds.

Why it matters

For end users on Android using Gboard, the underlying model update arrives automatically through Google Play Services and Gboard's own app update channel. Open the Play Store, search for Gboard, and tap Update if one is pending. You do not need to do anything special to activate Rambler or the improved transcription — the feature is on by default once the updated app is installed. To confirm Rambler is active, open any text field, tap the microphone icon in Gboard, and look for the Rambler option in the interface.

For Chrome users, the capability is rolling out as Google integrates Gemini 3.5 Transcribe into the browser. Make sure Chrome is updated to the latest stable release by navigating to the three-dot menu, selecting Help, and then About Google Chrome. The browser will check for updates automatically and prompt you to relaunch if one is ready. Once the feature reaches your channel, speech-to-text functionality in Chrome will route through the Gemini model without additional steps.

Who is affected

For developers accessing Google's AI or speech APIs, check the Google AI Studio dashboard and the Google Cloud Speech-to-Text documentation for any newly listed model identifiers referencing Gemini 3.5 Transcribe. If you have existing integrations pointing to a generic speech model, review the API reference to see whether a model parameter now accepts a Gemini-family identifier. Update your configuration to specify the model explicitly once it appears in the available options for your account tier.

The rollout is not instantaneous across all surfaces. Chrome integration is confirmed as coming but may not be present in the stable channel for all users at the time of this writing. If you do not see updated transcription behavior in Chrome, check whether you are on the latest release and whether the feature has reached your region. Android users on older versions of Gboard may not receive the update immediately if their device does not support automatic background updates from the Play Store. Manually checking for updates resolves this in most cases.

What to watch next

Developers should not assume that the Gemini 3.5 Transcribe model identifier is available in every API product or pricing tier at launch. Google typically stages access to new model versions. If the identifier does not appear in your project's model list, watch the Google AI changelog and the Cloud release notes for confirmation of general availability.

The expansion of Gemini 3.5 Transcribe beyond Gboard suggests Google is building transcription as a shared service layer. Watch for announcements about additional Google apps adopting it, further API access details for developers who want to call it directly from their own products, and any quality or latency comparisons Google publishes as the model reaches new surfaces. The Chrome integration in particular is worth tracking because it could enable browser-native transcription that works consistently across web applications without plugins.