Gemini Live may let users switch between speaking and typing without leaving the same conversation. Code in Google app version 17.58.13.sa.arm64 points to a work-in-progress design that would open a keyboard during an active Live session, accept a typed message, and let the user return to voice input through the microphone icon. The proposed mode remains a reported development rather than a public Gemini Live feature.
Gemini Live may add keyboard input, but the feature is not public
Gemini Live is built around back-and-forth voice conversations. The reported change would add a second input method inside that same experience instead of forcing users to leave Live whenever speaking is inconvenient.
The design could be useful in places where saying a prompt aloud is impractical, such as a quiet room or a noisy street. It would also make longer or more precise prompts easier to enter without breaking the conversation flow.
How the reported voice-and-text mode would work
If Google ships the reported design, the interaction would follow this sequence:
- Swipe up during a Live session. The gesture would open the Android keyboard.
- Type a message. Voice input would be disabled by default, with the microphone shown as crossed out while typing.
- Submit the prompt. The keyboard would hide after submission so Gemini Live could display its response.
- Tap the microphone icon to speak again. Users could reportedly move back and forth between typing and talking during the same conversation.
That is a meaningful change from treating text and voice as separate destinations. The proposed interface would keep both controls available within Live, while still making one input mode active at a time.
What Gemini already supports officially
Google’s documented workflow is narrower: users can type in a standard Gemini chat and then enter a Live voice conversation while retaining the thread’s context, tools, and conversation history.
That means Gemini already supports a handoff from text chat to Live voice chat. The reported keyboard mode would go further by allowing text entry after the Live conversation has already started.
For now, the official text-to-Live handoff is the available workflow, while keyboard input inside an active Gemini Live session remains a proposed feature.
