Google brings real-time video conversations to Search with Search Live

Troubleshooting a tangled mess of cables behind your TV used to mean squinting at serial numbers, Googling user manuals, and hoping you bought the right HDMI adapter. Google wants to turn that whole headache into a simple video call with search.

Google has officially rolled out Search Live in English to all users in the United States. Best of all, it is no longer locked behind an experimental waitlist—no Search Labs opt-in required.

Search Live lets you have a real-time, interactive voice conversation in AI Mode while sharing your smartphone camera feed. In short: Google Search can now see what your lens sees, answer spoken questions on the fly, and pull up relevant web links if you need to dig deeper.

Searching with your eyes and ears

Getting started takes just a tap inside the Google app on iOS or Android. You will find a new Live icon directly under the search bar. Tapping it lets you start speaking aloud and flips on video streaming for visual context.

If you are already inside Google Lens, there is also a new Live button resting at the bottom of the viewfinder. Launching it from Lens enables camera sharing by default, letting you immediately start asking questions about whatever is sitting in front of you.

Google highlighted five practical use cases for everyday life, ranging from travel navigation to DIY tasks. When setting up a home theater, you can simply aim your camera at the back of the set and ask which cable goes where, skipping the tedious step of typing model numbers.

Other examples include learning a new hobby, like aiming your camera at traditional matcha gear to identify each tool, or diagnosing kitchen substitutes. You can also point the camera at a stack of board games on vacation to ask which title fits your group, or use it as a real-time science tutor while watching home experiments bubble up on the counter.

Why it matters

Google's smartest move here is integrating real-time visual reasoning straight into its core search product, rather than cordoning it off inside a standalone, experimental chatbot.

For everyday users, describing physical objects—like an odd screw, an obscure cable port, or a kitchen utensil—has always been the hardest part of searching. Being able to just point, talk, and interrupt naturally eliminates that friction completely. The friction, for now, is simply whether people feel comfortable holding up their phone and talking to an AI in public.

For the tech landscape, this puts intense pressure on OpenAI and Apple. While OpenAI pioneered seamless real-time voice and vision demos, Google is bringing the technology directly to hundreds of millions of existing mobile searchers, critically paired with live web links. That link integration ensures Google protects its core traffic engine instead of replacing it with a closed conversational loop.

The real test will be reliability. Video streaming inference is notoriously expensive, and visual hallucination carries real risks. If Search Live misidentifies an electrical port or gives flawed DIY repair steps, users will lose trust fast. But if the latency holds up, searching by typing words into a box is going to feel ancient very soon.