How to use Agent in Google Flow

How to use Agent in Google Flow | Find Your Flow

Google Flow Agent is your creative partner that can plan and reason through complex tasks with your inputs, under your control. Built with Gemini models, it brings expertise and a deep understanding of your project to help with early brainstorming, creating and editing. For example, your agent can be a sounding board for dialogue between characters in a specific scene, or even make plot recommendations.

00:00 Introduction to Google Flow

00:32 Activating the Creative Agent

00:59 Prompting Variations & Character Adherence

01:26 Editing Assets & Organizing Collections with Gemini Omni

The Google Flow Agent is an AI creative assistant that streamlines brainstorming, asset generation, editing, and project organization. By turning it on in the prompt bar, the agent takes over manual settings and handles complex generative prompts. [1, 2, 3]

Getting Started

Enable the Agent: Open or create a project in Google Flow. In the prompt box, click the Agent icon to toggle it on. [1, 2, 3]
Set Agent Instructions: Before generating, click Agent Instructions to establish style guidelines, character sheets, or camera rules for consistency. [1, 2]
Configure Settings: Click the Settings icon at the bottom of the prompt box. Here you can set default models, aspect ratios, and turn on Confirm before generating so the AI seeks approval before burning your credits. [1]

What the Agent Can Do

Brainstorming: Use the agent as a conversational sounding board to outline storyboards or mood boards. [1, 2]
Generating Media: Describe your scene (subject, lighting, environment). You can drag existing media into the prompt box to provide visual context. [1, 2]
Editing & Batching: Drag existing videos or images into the prompt box and type what needs changing. You can also ask it to create multiple variations at once. [1, 2, 3, 4]
Organization: Ask the agent to rename specific files, group media into collections, or archive unused assets. [1, 2]

Could you tell me what specific task you are working on? Knowing if you want to brainstorm a storyboard, generate new video assets, or edit existing clips will let me give you precise, prompt-by-prompt instructions.

How to use Agent in Google Flow | Find Your Flow

The Google Flow Agent acts as your creative collaborator, powered by Gemini models to help you brainstorm, generate media, and manage your projects. It is designed to be a sounding board that understands the context of your specific work.

How to Activate and Use the Agent

Activation: Inside the prompt bar of any project, click the Agent button to toggle it on. The prompt box will change appearance, indicating that the agent is ready.
Expanding the Interface: For a more comprehensive workspace, click the Expand button to open the Agent side panel. This allows you to manage sessions and view chat history.
Deactivation: Simply click the Agent button again to turn it off and return to the standard prompt box.

Key Capabilities

Brainstorming & Planning: Use the agent to outline storyboards, develop mood boards, or turn abstract concepts into actionable prompts.
Media Generation: You can ask the agent directly to generate videos or images. You can also specify the model you wish to use.
Editing & Batch Processing: You can ask the agent to edit selected assets or create multiple variations (batch generation) of an asset simultaneously.
Context & References: Enhance the agent’s understanding by dragging media from your device or project directly into the prompt box. You can “tag” specific assets or characters in your prompts to maintain consistency throughout your project.
Organization: The agent can help you manage your workspace, such as renaming files, grouping assets into collections, or archiving unused items.

Customizing the Experience

Agent Instructions: To ensure consistent behavior, you can add specific guidelines by clicking Agent Instructions at the bottom of the prompt box. Here, you can upload reference images and provide descriptive guidelines for the agent to follow.
Settings: Click the Settings icon in the prompt box to configure defaults, such as:
- Confirmation: Toggle whether the agent asks for your approval before performing actions that consume AI credits.
- Generation Defaults: Pre-select your preferred aspect ratios, number of outputs, and the specific image or video models you want to use.

Managing Your Work

Sessions: Conversations with the agent are saved as “Sessions” specific to the project you are working in. You can access, rename, or delete these sessions via the Menu in the expanded Agent side panel.
Credits: Keep in mind that media generation performed by the agent consumes your Google Flow credits. While interacting with the agent for brainstorming usually does not cost credits, specific generation tasks will.

For further creative exploration or to see examples of what you can build, you can visit the Google Flow portal.

. This Is The First Real Shape Of AGI: Fusion Agents

. Your biggest ideas now bigger | June ’26 Pixel Drop

. Enhanced security monitoring with expanded Admin password reset alerts

. Join Google Meet calls from Safari on iOS devices

. Google Apps Script is now a Google Workspace core service with enterprise-grade data protection

. Connect to Google Meet hardware with room codes now in Early Preview

. Cannes Lions 2026: Strengthen creative campaigns with new tools from YouTube

. Tag Claude in, right where you already work

. World Surf League x Apple: How Apple Watch Changed Pro Surfing

. Updates to Gemini in Google Classroom

. Stricter classifications for Google Groups to enhance data security and privacy