Google Just Released “Flow” — And It Completely Destroys OpenAI’s Sora.

Google Just Quietly Released 2 Game-Changing AI Creative Tools (And Nobody Is Talking About It)

Last week, Google quietly dropped two of the most powerful creative tools on the internet, and almost nobody understands what they actually do yet.

🚨 WANTED: TECH TALENT

💰 High Pay | 🌍 Remote | ⚡ Fast Hiring

Frontend • Backend • Full Stack • AI/ML • DevOps

**👉 APPLY NOW**

In 2022, ChatGPT completely reshaped the writing landscape. Right now, Google is about to do the exact same thing to filmmaking, marketing, and content creation. They have released a powerhouse multimodal model called Omni, and built an AI-driven creative studio right on top of it called Google Flow.

Instead of just typing standard text prompts and hoping for a usable video clip, you can now feed Omni a mix of text, audio, images, or video references, and it will build exactly what you imagine. Once your video is generated, you can keep editing it just by talking to it in plain English.

Whether you’re a content creator wanting to look 10x more cinematic or a small business owner looking to scale your marketing, here is your ultimate, step-by-step beginner’s guide to mastering Google Flow before everyone else catches on.

Getting Started: The Google Flow Universe

You don’t need a high-end computer or any complicated installations to use this. Simply head over to labs.google/flow, sign in with your Google account, and you are inside the studio.

Before diving into a blank canvas, look at the top of your home screen for an icon called Google Flow TV.

The Ultimate Prompting Hack: Flow TV

Flow TV is a community library of AI-generated short films and clips made by other creators. The massive workflow trick here? Every single video comes with the exact text prompt used to generate it sitting right underneath the clip. If you want a specific “luxury” or “cyberpunk” aesthetic, search for it on Flow TV, copy the prompt that produced your favorite look, and use it as your starting baseline.

Tutorial 1: The “Locking” Workflow (The Cinematic Cartier Film)

To build a movie from scratch, hit New Project. Let’s break down how to use Flow’s built-in AI Agent — which acts as your personal Creative Director — to build a 75-second luxury jewelry commercial inspired by Cartier’s panther emblem.

Step 1: Briefing Your AI Director

Click the plus button at the bottom of the screen labeled Agent and feed it a structured brief.

The Brief: “Act as my creative director for a 75-second luxury jewelry film inspired by Cartier’s Panther emblem. Break it into a six-shot storyboard where a panther cub explores an empty library and discovers a glowing jewelry box. For each shot, give me the setting, camera move, lighting, and mood. Keep the whole thing warm and intimate.”

Before hitting enter, adjust your Agent Settings:

• Set Confirm before generating to Always (so you don’t accidentally burn credits).
• Set aspect ratio to 16x9 (YouTube friendly).
• Select Nano Banana 2 for images and Omni Flash for video generation.

Step 2: The Golden Rule of AI Consistency

The Agent will return a beautiful 6-shot storyboard layout. But do not blindly generate all the shots at once. If your main character or your background shifts slightly between shots, the cinematic illusion breaks.

Always follow this order:

1. Lock the Character First: Tell the agent to generate the panther cub. Pick your favorite render (or tell it to blend the aesthetics of two renders).
2. Lock the Location Next: Generate and approve the Parisian library setting.
3. Generate the Storyboard Grid: Once character and location are locked, the model anchors them. The resulting storyboard will feature the exact same cub in the exact same library across all panels.

Step 3: Directing and Assembling the Timeline

Approve the shots to let the Agent generate the actual video clips. If a shot isn’t quite right, use the Describe your edits text box directly underneath the clip (e.g., “Reframe this as a slow push-in”).

Finally, head over to the Scenes tab on the left sidebar to open your editing timeline. Drag and drop your six clips in order, hit play, and watch your cohesive, 60-second commercial come to life. You can change the title to “Cartier” and download the entire asset pack as a zip file.

Tutorial 2: The Manual Character Workflow (The Cadbury-Style Ad)

You don’t always need the AI Agent holding your hand. For full control, you can use the Characters tab to build recurring actors for commercial marketing. Let’s create a classic, emotional Cadbury-style ad.

Step 1: The Product Shot

Set your model to Nano Banana 2 and generate a macro product shot of a glossy chocolate bar breaking apart. Flow lets you generate this across multiple aspect ratios simultaneously (9x16 for Reels, 1x1 for Instagram Feed, 16x9 for YouTube) in under a minute, giving you a ready-to-go multiplatform asset pack.

Step 2: Creating Reusable Characters

Go to the Characters tab on the left sidebar:

• Create Character 1 (The Child): Prompt a photorealistic 8-year-old with natural styling and a genuine smile.
• Assign a Voice: Flow allows you to select a permanent voice profile for your character. This ensures their voice stays identical across dozens of future video episodes without drifting.
• Create Character 2 (The Neighbor): Repeat the process for an elderly neighbor.

Step 3: Baking Ingredients Into Your Prompt

Switch your generation engine from Image to Video and choose Omni Flash. Write your scene prompt (“A child on a stoop snaps a chocolate square and offers it to an elderly neighbor”).

Before hitting generate, click Add Ingredient and select your saved Child and Neighbor characters from your library. This forces the model to use your exact assets, guaranteeing perfect commercial consistency.

Unleashing the Tools Section: Custom AI Workflows

Flow isn’t just a video generator; it’s a creative operating system. The Tools tab contains powerful pre-built engines:

• Scene Explorer: Type in a location (like “1980s Taj Mahal”) and select multiple camera tags (wide shot, low angle, high angle). It will instantly render the exact same scene from those distinct cinematic perspectives.
• Style Switcher: Drag and drop an asset — like your photorealistic child character — and apply a target style like Studio Ghibli. It generates an anime version with a real-time slider to compare the two.

The Pro Move: You can build and save your own custom tools. If your brand uses a specific color grading or asset template, build it once, save it under My Tools, and turn a repeated creative task into a one-click button forever.

The Secret Formula: DeepMind’s Official Omni Prompting Guide

Google DeepMind published the official prompting manual for the Gemini Omni model. If you want Hollywood-grade outputs instead of mediocre AI fluff, your prompts should always cover these 5 Core Building Blocks:

[Camera Framing & Motion] + [Visual Style] + [Lighting Setups] + [Location Environment] + [Core Action]

• Shot Framing & Motion: Explicitly state if you want a dolly zoom, a slow push-in, or a static, locked-off camera.
• Style: Keep it simple. Specify “cinematic,” “photorealistic,” or “watercolor.” Don’t overengineer it; let Omni’s native reasoning fill in the gaps.
• Lighting (The Unsung Hero): Tell Omni where the light source is (e.g., sunlight through a window, a neon street lamp off-screen) and the mood it should create (ethereal, harsh, warm).
• Location: Use broad strokes. Omni knows what a “Grand Parisian Library” looks like; you don’t need to describe every single brick.
• Action & Text: Clearly define what the subjects are doing. If you want text overlays, specify the placement and animation style — Omni will render text in perfect sync with the audio beat.

How to Automate This with Claude

To make this effortless, copy the URL of Google’s official Omni prompting guide and paste it into Claude.

Create a permanent Skill (or Project Custom Instruction) titled “Gemini Omni Prompt Optimization Guide.” Instruct Claude to always cross-reference DeepMind’s rules whenever you ask for a prompt. Now, anytime you type a simple idea (like “Twinkle Twinkle Little Star”), Claude will instantly output a fully structured, Omni-optimized prompt ready to paste into Flow or the regular Gemini App (with Extended Thinking enabled).

Don’t Let the AI Gap Catch You Behind

90% of people read guides like this, nod along, and close the tab. Six months from now, they wonder why competing brands and creators are suddenly lightyears ahead of them.

The window opened this week. The gap between the creators who jump into tools like Google Flow in week one versus those who wait half a year will be the most expensive gap in the digital industry this year.

Open a browser tab, log into Flow tonight, and build just one 30-second clip. Your creative workflow will never be the same.

Thank you for being a part of the community

Before you go:

👉 Be sure to clap and follow the writer ️👏️️

👉 Follow us: **Linkedin| [Medium](https://medium.com/codetodeploy)**

👉 CodeToDeploy Tech Community is live on Discord — **Join now!**

Disclosure: This post includes affiliate and partnership links.