Talking to Gemini Live used to mean starting over every single time. No matter how many details you shared in a previous session, the next conversation began with a clean slate. That changes now: Google has quietly extended memory access to Gemini’s conversational mode, so the assistant can pull relevant details from past chats when answering new questions.
The shift was first spotted by 9to5Google on a Google support page, which describes Gemini Live gaining access to memory of past conversations along with information from select Connected Apps. In practice, that means the things you mention once — dietary restrictions, an allergy, an important family date — can now stick around, sparing you the tedium of repeating yourself each time you fire up voice mode.
Closing an awkward gap
Memory has been part of Gemini’s standard text chat for over a year, but Live operated without it, creating a strange divide depending on whether you typed or spoke. Ask a written question and Gemini remembered the context; ask the same thing out loud and it drew a blank. This update finally lines the two experiences up.
One inconsistency lingers, though. According to 9to5Google, the Personal Intelligence settings page on Android still lists memory as “coming soon” for Live, even though the feature already works for some users. That mismatch points to a staggered rollout, so it may not have reached your device just yet.
Here’s what we know about availability:
- Memory for Gemini Live is rolling out in English in the US for now
- The capability is included in the free Gemini tier, where Live voice mode is available
- The update appeared on a Google support page after May 23, 2026
- No word yet on wider language support or broader regional availability
It’s worth remembering that memory is the kind of feature you only notice when it’s absent. Flashier additions grab headlines, but recall quietly does more for everyday usefulness — and rivals had already crossed this line. ChatGPT’s voice mode picked up memory before Google’s, so this is as much about catching up as it is about innovation.
If the rollout mirrors how memory debuted in standard Gemini, broader language support and wider availability should arrive in stages over the following months. For now, US English speakers are the test bed, and the early signs suggest a smoother, less repetitive back-and-forth with the assistant that finally stops forgetting who you are between sessions.
