Gemini Live
Last reviewed
Jun 3, 2026
Sources
13 citations
Review status
Source-backed
Revision
v1 · 1,491 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 3, 2026
Sources
13 citations
Review status
Source-backed
Revision
v1 · 1,491 words
Add missing citations, update stale details, or suggest a clearer explanation.
Gemini Live is a conversational voice feature in Google's Gemini app that lets people hold free-flowing spoken conversations with the company's AI assistant. Rather than typing prompts or issuing one-off voice commands, a user taps a dedicated control and talks to Gemini much as they would on a phone call, with the ability to interrupt, change topics, and pause mid-conversation. Google introduced the feature in August 2024 and steadily widened both who can use it and what it can do, eventually adding real-time camera and screen-sharing input drawn from its Project Astra research.[1][2]
When Google launched the standalone Gemini app, its voice support worked in a turn-based fashion similar to traditional digital assistants: a user spoke a request, the assistant answered, and the exchange ended. Gemini Live was positioned as a more natural alternative built around continuous, two-way dialogue. Google framed it as a way to brainstorm out loud, talk through ideas, or get help hands-free while doing something else, with the assistant designed to handle the rhythm of ordinary human speech rather than discrete commands.[1][3]
The feature first appeared alongside the Pixel 9 phones at Google's annual hardware showcase, where Gemini more broadly was being promoted as a replacement for Google Assistant on the newest devices. Gemini Live was one of the marquee additions used to demonstrate how the assistant could feel more like talking to a person.[2][4]
Google announced Gemini Live on August 13, 2024, at its Made by Google event. At launch it was a mobile experience limited to Gemini Advanced subscribers, who paid for it through the Google One AI Premium plan, and it was available only in English and only on Android phones. Google said at the time that the feature would expand to iOS and to more languages in the following weeks.[1][2]
The paywall on the core voice feature did not last long. By early October 2024 Google began offering Gemini Live to all users on Android at no charge, dropping the requirement for a paid subscription, though it remained English-only at that stage and some adjacent capabilities, such as deeper integration with Gmail and Docs, stayed tied to the premium tier.[5][6]
The next major expansion came in 2025 with visual input. Project Astra-powered camera and screen sharing began rolling out to Gemini Advanced subscribers on Android starting around March 22, 2025. Google indicated that owners of recent Pixel and Galaxy S25 phones would be among the first to receive these capabilities, although the rollout reached a range of Android devices over the following weeks.[7][8] In April 2025, Google removed the subscription requirement for camera and screen sharing on Android, making them free for all Android users.[9]
At Google I/O in May 2025, the company extended the visual features further. It announced that Gemini Live with camera and screen sharing would be available to everyone on both Android and iOS for free, with the iOS rollout noted as starting on May 20, 2025.[10] Coverage of the iPhone rollout confirmed wider availability later in May, reaching both free and paid accounts in the United States and requiring the latest version of the app.[11]
The table below summarizes the main milestones.
| Date | Milestone |
|---|---|
| August 13, 2024 | Announced at Made by Google; voice feature limited to Gemini Advanced, Android, English |
| October 2024 | Voice feature made free for all Android users |
| March 22, 2025 | Project Astra camera and screen sharing begin rolling out to Gemini Advanced on Android |
| April 2025 | Camera and screen sharing made free for all Android users |
| May 2025 (I/O) | Camera and screen sharing made free on Android and iOS; iOS rollout from May 20 |
| May 29, 2025 | Wider iPhone rollout of camera and screen sharing in the US |
| August 2025 | More expressive audio model and Google app connections added |
A user opens Gemini Live by tapping a dedicated Live control in the Gemini app and then simply speaks. The defining trait is the conversational flow: a person can interrupt the assistant mid-response to add detail or steer the discussion, pause a chat and return to it later, and follow up without restating context. Google's own description emphasizes that you can talk, pause, interrupt, pivot, and follow up much as you would with another person.[3][12]
The feature is built for hands-free use. Conversations can continue while the Gemini app runs in the background or while the phone is locked, so a user can keep talking on the move in a manner Google likened to an ordinary phone call.[1][3]
Voice is a central part of the experience. At launch Google offered 10 voice options so users could pick a tone and style they preferred.[1] The selectable voices carry distinct names, including Vega, Ursa, Pegasus, Nova, Eclipse, Lyra, Orbit, Dipper, Capella, and Orion.[12] A user can switch a single ongoing conversation between the voice (Live) mode and the standard text mode without starting over.[12]
Beyond voice, Gemini Live gained the ability to take in real-time visual information. With the camera feature, a user can point a phone at the world and ask questions about what is in view, and the assistant uses that live context to answer. With screen sharing, a user can broadcast the contents of their phone screen and ask Gemini about whatever app or page is open. In both cases the visual stream stops automatically when the user leaves the app or locks the screen, while the spoken conversation can continue.[7][11]
These visual capabilities trace directly to Project Astra, Google DeepMind's research effort on a universal AI assistant that can perceive and reason about its surroundings in real time. Google explicitly described the rollout as powered by Project Astra.[7][11] After the 2025 expansions, the camera and screen-sharing features moved from being a paid, Android-only perk to something available free across Android and iOS.[9][10]
Gemini Live is powered by Google's Gemini family of models, the same line of multimodal systems that underpins the Gemini assistant more broadly. The feature relies on the models' ability to handle long, continuous exchanges; early coverage noted that Gemini Live could draw on a long context window to sustain extended back-and-forth dialogue.[2] Real-time, low-latency interaction of the kind Gemini Live depends on aligns with the streaming, multimodal direction of models such as Gemini 2.0 Flash and the Project Astra prototypes that fed into the camera and screen-sharing features.[8][7]
Google continued to upgrade the experience after launch. In August 2025 it rolled out audio model improvements aimed at making the assistant sound more natural, giving it better control over intonation, rhythm, and pitch, and letting users ask it to speak more slowly, more quickly, or in a particular accent. The same update connected Gemini Live to more Google apps, including Calendar, Keep, Tasks, Messages, and Maps, and brought visual guidance to the Pixel 10 series before expanding to other devices.[13]
Press coverage generally treated Gemini Live as Google's answer to other conversational voice assistants, with much of the commentary focused on how natural the back-and-forth felt and on the steady removal of paywalls. Outlets noted that making the voice feature, and later the camera and screen-sharing tools, free for all users marked a notable shift in Google's strategy, putting capabilities that had started as premium features into the hands of a much broader audience.[6][9] Reporting also tracked the rollout closely, in part because availability arrived gradually and varied by device and region rather than appearing for everyone at once.[8][11]