Google Whisk: Convert images to talking AI videos with Google Flo

Google Whisk Try it →Watch full video →Learn AI with Ritika · Nov 2025

Demo summary

The creator uses Google Flo AI (incorporating the Vidu 1.5 model) to upload images and JSON prompts from ChatGPT to generate video clips with synchronized dialogue and background music.

Step-by-step

Log in to Google Flo using your email ID.
Click on New Project and select the Frames to Video option.
Use ChatGPT to convert your scene description and dialogue into a JSON video prompt for Vidu 1.5.
Copy the generated JSON code from ChatGPT.
Paste the JSON code into the prompt box in Google Flo.
Click the plus (+) sign to upload your base image.
Select the Aspect Ratio (Portrait or Landscape).
Click the white arrow button to generate the video.
Preview the generated options and click the download arrow on your preferred version.

Options

Convert Text to Video
Convert Ingredients to Video
Choose between Portrait or Landscape aspect ratios

Watch out for

Google Flo requires specific JSON prompts to generate videos with synchronized dialogue and background music.

Tips

Use ChatGPT to describe the character's actions, such as waving or adjusting a camera, to make the scene more dynamic.
You can use the creator's provided ChatGPT prompt template as a base for your own scenes.
Google Flo typically generates two video options for each prompt so you can choose the best result.

Image to Video

4:512:48Convert images to talking AI videos with Google FloCurrentThe creator uses Google Flo AI (incorporating the Vidu 1.5 model) to upload images and JSON prompts from ChatGPT to generate video clips with synchronized dialogue and background music.Learn AI with Ritika
9:001:53Animate images with Whisk AIThe user demonstrates generating an image from a prompt in Whisk AI and then using the animation icon to transform that static image into a video.AI Playground
2:030:52Generate image-to-video with audio in Google Veo 3The creator demonstrates using Google Flow to access Veo 3, uploading an image from Open Art, and entering a prompt to generate a video with synchronized audio and dialogue.Dan Kieft
10:360:39Create a consistent character in Google FlowThe creator demonstrates how to upload a reference image of himself to Google Flow, name the character, and use an '@' mention in a prompt to generate a video of that specific person winning a race.AI Samson
7:080:42Image-to-video with Google Veo 3The user demonstrates the 'frames to video' feature in Google Flow by uploading a jungle scene image and providing a complex prompt to animate specific animals within the frame.Tim Harris AI
3:510:26Animate Images with Whisk AIThe creator demonstrates the 'Animate' feature in Whisk AI, showing how to add a video prompt to a generated image to create a motion clip.Novid | Automation