Google’s Flow Just Learned New Tricks, and Video Editing Will Never Be the Same

Google Labs has rolled out a fresh batch of upgrades for Flow, its AI-powered video tool that has already churned out more than 500 million clips since launching in May.

The message behind the update is clear. Flow is no longer a quirky experiment. It is quietly becoming one of Google’s most aggressive plays in creator tech.

The headline addition is image generation powered by something delightfully named Nano Banana Pro. The upgrade lets users reshape scenes without restarting the entire clip.

A different outfit, a new pose, a moodier light source, a fresh camera angle. Everything can be swapped with a few prompt tweaks.

Product manager Anika Ahluwalia says the system is designed to let people refine shots with minimal friction, rather than re-rolling the whole sequence every time they want a change.

Google is also leaning into more tactile editing. Flow now supports doodle-based prompting, which means you can literally scribble over a frame to guide the model.

Need a prop moved, or a background element shifted? Draw it. The tool interprets your lines and adjusts the image accordingly.

One of the more anticipated features is object insertion and removal. Flow will let users drop items in or erase them entirely. Object removal will arrive as an experimental option next month, which is Google-speak for “it might break but please try it anyway.”

A new reshoot feature rounds out the update. It gives creators control over camera motion inside generated videos, allowing them to reposition or rethink shots without going back to square one.

Google Labs says the goal is to hand creators finer control and to push what AI-assisted editing can do.

The subtext is obvious. With rivals racing to build their own generative video engines, Flow needs to evolve fast. This update shows it is doing exactly that.

Is this conversation helpful so far?