Google has unveiled Search Live, a voice-powered conversational interface built into AI Mode in the Google mobile app for Android and iOS, currently rolling out to U.S. testers signed up via Google Labs . Its goal: to transform mobile search into a natural, back‑and‑forth dialogue with Google’s Gemini 2.x AI model.
Unlike traditional voice search, which treats each query independently, Search Live enables multi-turn conversations—you speak, Google responds verbally, and you can follow up right away without retyping. Google’s design supports users on the go or multitasking, with the app continuing to listen even when you’re switching apps .
How It Works
- Opt-in via Labs: To enable, users in the U.S. must join the “AI Mode” experiment in Google’s Search Labs .
- Launch Live Mode: Open the Google app, tap the new Live (“waveform”) icon under the search bar .
- Ask by Voice: Pose questions like, “How do I prevent linen clothes from wrinkling in a suitcase?” and the system replies audibly via Gemini-based voice .
- Multi-turn Dialogues: You can quickly ask follow-ups such as: “What if they still wrinkle?” — and Google will respond seamlessly, retaining conversational context .
- On-Screen Links: Even while talking, a carousel of relevant search links appears to allow manual follow-through .
- Transcript Viewing: Tapping “Transcript” converts the spoken response into text—letting you switch between voice and text interactions .
- History Saved: Complete dialogue history is stored in your AI Mode history, letting you revisit or build on previous conversations .
- Background Operation: The voice interaction continues even with other apps open or the device locked .
🤖 Behind the Scenes: Gemini + Query Fan-Out
“Search Live in AI Mode uses a custom version of Gemini with advanced voice capabilities,” said Liza Ma, Director of Product Management at Google Search . This custom Gemini is augmented by Google’s robust search infrastructure using a technique called query fan-out, which allows the system to simultaneously draw information from a variety of web sources, resulting in broader and more nuanced answers .
Why It Matters
Hands-free convenience: Perfect for tasks like cooking, driving, or packing. You can engage in useful searches without typing .
Natural conversation flow: The system maintains the dialogue context so you don’t have to re-explain every nuance .
Verbal + visual synergy: Audio responses paired with clickable links offer both quick helpful info and deeper exploration .
Future multimodal possibilities: Google plans to soon integrate camera-based live visual search, so you’ll be able to show something to the AI and discuss it in real time .
Context: A Competitive Landscape
Search Live arrives amid a surge in voice-powered AI assistants:
OpenAI launched ChatGPT Advanced Voice Mode last year.
Anthropic added voice to its Claude AI in May 2025.
Apple is working on an “LLM Siri,” though rollout was delayed to ensure reliability .
Google’s search-first approach—rooted in Gemini and backed by the web’s depth—gives it a distinctive edge compared to standalone chatbots .
Limitations & Beta Stage
Limited availability: Only U.S. users who opted in via Labs have access.
Beta voice quality: The AI voice may still sound slightly robotic .
Privacy considerations: As always with voice search, there may be implications. Google hasn’t yet detailed storage or duration of voice transcripts.
Multimodal coming later: While projector Astra for live camera interactions is mentioned, that integration is still forthcoming .
What’s Coming Next
Google has hinted at future enhancements:
Live multimodal search: The ability to show your camera view and ask questions—similar to Gemini Live .
Improved integrations: Features like Project Mariner will enable deeper web-page understanding, assisting with tasks like shopping or bookings .
Visual search convergence: Expect seamless switching between voice and camera inputs for richer real-time dialogues .
Conclusion
Google’s Search Live in AI Mode marks a significant shift in mobile search—moving from isolated voice queries to continuous, natural conversations powered by Gemini. While still early stage, it promises a hands-free, context-aware experience, anchored by web integration. As Google rolls out multimodal capabilities, Search Live could redefine how we interact with our phones—blurring lines between search, voice assistant, and real-world understanding.
For U.S. testers in Google Labs, now’s the perfect time to opt in and explore. And if you’re outside the U.S., stay tuned—this feels like a global launch waiting to happen.
