You can now use Veo 2, Google’s AI-powered video generation model, on Google AI Studio. It supports text-to-image and, more importantly, image-to-video. You can generate a high-resolution (720p) video clip for up to 8 seconds using a simple prompt or optionally with an image.
If you use online AI video services like Kling or Runway ML, Veo 2 is a nice addition to your toolbox.
In this article, you will learn what the Veo 2 model is and how to generate video with it.
Detailed voxel art illustration with large blocks. A voxel hand holds a melting voxel ice cream cone against a vast, clear blue sky with a bright voxel sun. The ice cream shows is melting, with voxelized drips and droplets falling down. The style is consistent, colorful, blocky voxel art throughout the hand, ice cream, drips, and sky. Time-lapse concept. Focus on the ice cream. Zoom in on the ice cream. pink, red, blue ice cream scoops.A sleek black Lamborghini speeds through a lush tropical rainforest, sunlight filtering through dense emerald canopies, leaves fluttering in its wake. The glossy car reflects the vibrant green surroundings as it glides over a winding jungle road, mist rising from recent rain, with birds flying overhead and distant waterfalls roaring in the background. Ultra-smooth tracking shots, cinematic lighting, and dynamic motion blur create a high-end, visually stunning scene — futuristic yet natural.A serene lake nestled deep within a lush, mystical jungle. Vivid green canopies sway gently above, casting dappled light on the shimmering water. Exotic, tropical birds with iridescent feathers soar gracefully through the air, their calls echoing in the stillness. Giant, otherworldly flowers bloom along the banks, glowing faintly with magical energy. Ethereal mist rolls across the water’s surface, as if the lake itself is alive with ancient secrets. In the distance, a crystal waterfall cascades from towering cliffs, sparkling with light from an unseen sun. Everything feels enchanted — a hidden paradise untouched by time.
Veo 2 video model
Google hasn’t disclosed the model architecture.
Videos generated by Veo 2 include an invisible SynthID watermark to curb misinformation, so people can easily identify if the content was created by AI.
Benefits:
Smoother video quality overall compared to other models
Drawbacks:
Video generation quota is around 5 per day.
However, this can be bypassed using another Google account
Google doesn’t save your videos for you, so you’ll have to download them after generation
Enable Google Drive Access by clicking on the prompt section, and start writing your prompt.
Step 4: Revise settings
Select the number of generations (1-2). Each video takes roughly 35 seconds to generate, so the wait time isn’t too bad, generating 2 videos.
Select preferred aspect ratio (16:9) or (9:16)
Select video duration (5 – 8 seconds)
Scroll down and write your negative prompt (Things you DO NOT want Veo 2 to generate)
Step 5: Set the initial frame (Optional)
Upload / Select an image for the starting frame.
Step 6: Generate a video
Click run to generate your video!
Generated with Van Gogh’s Starry Night as the image prompt, text prompt: The stars slowly begin to shimmer and swirl in motion. The village lights flicker gently, trees sway softly, and the entire painting breathes with a dreamlike, magical energy
By Andrew
Andrew is an experienced software engineer with a specialization in Machine Learning and Artificial Intelligence. He is passionate about programming, art, and education. He has a doctorate degree in engineering.
Thanks for alerting us to this, Andrew. It gives great results and as you say, it is very quick. I got a very nice 5 sec video with this prompt: A steampunk male character enters a steampunk laboratory and starts adjusting the controls of his time machine which stands in the middle of the room. A blue light above the machine starts pulsing as it begins to generate a time vortex. A retrofuturistic steampunk scene.
I struggled to get the platform to accept an image as a start frame, however. It says images with people aren’t supported but even even the ones without that I tried were blocked for “safety” reasons. So that element is more limited in scope than Kling at the moment, I think.
Thanks for alerting us to this, Andrew. It gives great results and as you say, it is very quick. I got a very nice 5 sec video with this prompt: A steampunk male character enters a steampunk laboratory and starts adjusting the controls of his time machine which stands in the middle of the room. A blue light above the machine starts pulsing as it begins to generate a time vortex. A retrofuturistic steampunk scene.
I struggled to get the platform to accept an image as a start frame, however. It says images with people aren’t supported but even even the ones without that I tried were blocked for “safety” reasons. So that element is more limited in scope than Kling at the moment, I think.