Create videos in Google Flow

You can animate images and create videos from text prompts, ingredients, frames, and other videos in Google Flow.

Learn how to edit videos and build scenes in Google Flow .

Create video clips in Google Flow

Create videos with the Google Flow Agent

1. On your device, go to Google Flow .
2. Click an existing project or start a new one.
3. In the prompt box, turn on Agent.
   • Learn how to use the Google Flow Agent .
4. Describe the scene you want to create in detail. Be specific about the subject, action, environment, lighting, and style.
   • You can ask the Agent to create multiple variations of a video at once.
   • For example:“Give me 5 variations of this video with different lighting.”
5. Optional: To add ingredients to your prompt, drag media into the prompt box from your device or project.
6. Click Generate.

Create videos with the standard Google Flow prompt box

1. On your device, go to Google Flow .
2. Click an existing project or start a new one.
3. In the prompt box, describe the scene you want to create in detail.
   • Be specific about the subject, action, environment, lighting, and style.
4. Click the model name.
   • By default, Google Flow sets the model to Nano Banana Pro.
5. Click Video.
6. Optional: Add ingredients or frames to your prompt.
   • Learn how to add references to your prompt .
   • Learn how to add start and end frames to your prompt .
7. Select your video preferences:
   • Aspect ratio.
   • Number of outputs you want.
   • The model you want to use. Learn more about Google Flow models and supported features .
   • Generation length.
8. Click Generate.

Add start & end frames to your prompt

Important:Product feature availability may differ by country/region of use. Certain image types may not be supported.

You can create transitions, animate images, or generate a video that starts and ends with specific images with the use of frames.

1. In your project's prompt box, click the model name.
2. Click Video Frames.
3. Add frames to your prompt.
   • To add your start frame:Drag an image to “+ Add start frame.”
   • To add your end frame:Drag an image to “+ Add end frame.”
4. In your text prompt, describe the action or transition that should happen between the frames.
   • You can also prompt Google Flow to create an animated image.
5. Select your video preferences.
6. Click Generate.

Add references to your prompt

Add characters to your prompt

1. On your device, go to Google Flow .
2. Click an existing project or start a new one.
3. In the prompt box, type @ followed by your character's name.
   • To add your avatar to your prompt, type @me . 
   • Learn how to create & use your avatar in Google Flow.
4. Describe how your character should be used as a reference.
   • For example: @CaptainZoro walking through a futuristic city. 

Tip: Learn how to create a character in Google Flow.

Add visual references to your prompt

You can consistently use the same images for your character and key objects from one clip to the next with the use of ingredients.

1. In your project’s prompt box, click the model name.
2. Click Video Ingredients.
3. Add your ingredients in any of these ways:
   • From your device, drag your reference image or video into Google Flow.
   • To search and select any uploaded asset in your project by name, in the prompt box, type @ .
   • To upload files from your computer or select uploaded assets in your project, under your text prompt, click Add.
4. In your text prompt, describe how your ingredients should be used as references within the clip.
   • Example:With ingredients of a woman, a lava lamp, and a foggy street, you can enter the prompt: The woman, whose torso is the lava lamp, walks down the foggy street .
5. Select your video preferences.
6. Click Generate.

Tip:For the best results, provide subject or product references on a plain or segmented background.

Add voice references to your prompt

To maintain a specific character’s voice across multiple video clips, add a single-speaker voice reference to your prompt.

1. In your project’s prompt box, click the model name Omni Flash .
2. Click Video Ingredients.
3. Under your text prompt, click Add Voices.
4. Choose your preferred voice.
   • To hear a 10-second sample preview of a voice, hover over it. Then click Play.
   • To reference your chosen voice directly, in your text prompt, type @Voice .
     • Example: @Voice: Andrew .
5. Select your video preferences.
6. Click Generate.

Create custom voices

1. In your project’s prompt box, click the model name Omni Flash.
2. Under your text prompt, click Add Voices Create New Voice.
3. Under “Base Voice,” select a preset voice.
4. Under “Name,” name your voice.
5. Under “Voice Performance,” describe how to change the voice.
   • For example:“Make the voice sound slightly raspy with a New York accent.”
6. Optional: To preview an 8-second sample of your custom voice, under “Sample Dialogue,” enter a short text string.
7. Click Sync.
8. Click Save New Voice.

Understand limitations & guardrails

You can add voice references only to video generations that use ingredients. For all other kinds of generations, you’ll get an error.

Best practices & advanced tips when generating in Google Flow

Get the most out of every generation with these tips:

• Prepare clean ingredients
  • When you add ingredients to your prompt, provide subject or product references on a plain or segmented background.
  • Make sure location and style references don't contain extra subjects, unless that is your creative intent.
• Avoid conflicting guidance
  • Your text prompt should complement, not contradict, your visual inputs.
  • Make sure your prompt references the frames or ingredients you've provided.
  • A consistent look and feel across all your ingredient images helps the model blend them more effectively.
• Iterate with Gemini
  • As a Google AI subscriber, you have access to the most capable Gemini models.
  • Visit gemini.google.com to help brainstorm. Send Gemini your prompt, an image, or even a video and ask it to write you new prompts, riff on ideas, and more.