AI video generator

Turn images or text into AI videos

Nano Banana makes still images, not video. Create or edit an image with Nano Banana, or bring your own photo, then animate it with a video model such as Veo 3, Seedance 2, Kling 3.0 or MiniMax H3. You can also start from a text prompt.

Image-to-video and text-to-video with Veo 3, Seedance 2, Kling 3.0 and MiniMax H3
Videos render in the background and finished clips appear in My Videos
Review model settings and the credit estimate before generating
Generate now

Turn an image into a video

Image to video takes four steps: choose Frames and upload your image as the first frame, describe the motion, pick the model, duration and resolution, then check the credit cost shown on the Create button and submit. The cost depends on the model and settings; plans and credit packs are listed on Pricing. For text to video, choose Text and describe the scene instead.

173 / 20000
X 1
Audio
Generate synchronized sound
Web search
Add online context
Details:1094 Credits
After you submit, generation continues in the background and finished videos appear in My Videos.

Sample Video

Seedance 2 Sample Video
Generated Clip
Cinematic product motion
Prompt, audio, camera, and ratio controlled in one run.
Duration
4-15s
Max quality
1080p
Audio
Native

Choose a video model by settings and task

Minimum-duration examples. Seedance 2 and MiniMax examples exclude optional video references and extra images. Motion Control uses measured motion-video duration.

Model / versionLowest settingSecondsResolutionAudioUseful for
Veo 3672 Credits / 4s · 720p4–8720p, 1080pYes, always onShort scenes with built-in sound
Veo 3 Fast160 Credits / 4s · 720p4–8720p, 1080pYes, always onShort scenes with built-in sound
Seedance 2254 Credits / 4s · 480p4–15480p, 720p, 1080pOptionalProduct or character clips guided by references
Seedance 2 Fast207 Credits / 4s · 480p4–15480p, 720pOptionalProduct or character clips guided by references
Seedance 2.5374 Credits / 4s · 480p4–30480p, 720p, 1080pOptionalProduct or character clips guided by references
Kling 3.0140 Credits / 3s · 720p3–15720p, 1080p, 4KOptionalPlanned shots or first/last-frame scenes
Kling 4.0285 Credits / 3s · 720p3–30720p, 1080pYes, always onProduct or character clips guided by references
MiniMax H3107 Credits / 4s · 768P4–15768P, 2KYes — observed 2026-10-23 in text/image tests at 768P, 4 seconds; no audio toggle.Reference-guided drafts
Kling 3.0 Motion Control200 Credits / 3s · 720p3–30s (video orientation); 3–10s (image orientation)720p, 1080pNot verified; inspect the resultTransfer motion from a reference video
Gemini Omni Flash

Omni generation is not available here; this page currently generates with Veo 3 Fast.

160 Credits / 4s · 720p4–8720p, 1080pYes, always onShort scenes with built-in sound

Pick by task

For a product ad with reference images, start with Seedance 2. For a planned sequence of shots, use Kling 3.0. For a short scene with sound, use Veo 3. These choices follow available controls, not a quality ranking.

Image to Image AI is independent of the model vendors. Upload only content you own or have permission to use. Review generated results before publishing.

Why it works

AI video generation for motion-first tasks

Plan one clear action per clip: a slow product reveal, a camera push toward a subject, or a subtle animated scene. A specific movement and a simple composition make the result easier to evaluate.

Built for short-form motion

Describe scenes, pacing, camera direction, and subject movement when a still image is not enough.

Reference images for visual continuity

Start from a product, character, or composition when the selected model supports image-to-video.

Runs in the background

Generation takes time. Follow the job status and review completed clips in My Videos before starting another variation.

Workflow

From prompt to generated video in four steps

Start with one short test, review the movement and visual consistency, then refine the prompt before making more versions.

1

Choose the video task

Start from text-to-video or add a reference image when the model supports image-to-video.

2

Describe the motion

Write the scene, subject, camera movement, pacing, style, and important constraints.

3

Select model settings

Pick the model, duration, quality, and aspect ratio available for the task.

4

Submit and review

Check the displayed credit estimate before submitting. Review the completed clip for subject consistency, unwanted movement and framing. If a job fails, check its status before submitting again.

Use cases

Short video ideas for creators and teams

Compare image models for your source frame

Product clips

Create short product motion, reveal shots, packaging demos, and ecommerce video drafts.

Creative ads

Generate campaign concepts, social video ideas, and visual directions for ad production.

Image animation

Turn selected visuals into motion with image-to-video.

Model testing

Compare output quality, speed, and credit cost before repeating video generation at scale.

Capabilities

Write a motion prompt you can test

Prompt example, not a generated result: “A ceramic mug on a wooden table beside a window. The camera slowly pushes in as steam rises. Soft morning light, stable framing, no cuts.” For image-to-video, upload your mug photo and focus the prompt on movement rather than redesigning the mug.

Text-to-video

Turn a prompt into a short generated video.

Image-to-video

Use references when the model supports image animation.

Motion prompting

Describe camera movement, subject action, and pacing.

Background generation

Follow generation status and review the result in My Videos.

Duration and quality

Choose available model settings before submitting the job.

Check before sharing

Play the full clip and check faces, product details, flicker and unexpected camera movement.

FAQ

Nano Banana AI video FAQ

How do I turn an image into a video?

Image-to-video animates a still image with a video model. Choose Frames, upload a photo you own or have permission to use, then describe the movement. For example: “Keep the product unchanged; slowly move the camera closer while the background stays still.” Check the selected model’s duration, resolution and credit estimate before generating, then review the result for unwanted changes.

Can I use image-to-video AI with free credits?

Yes, if your account has enough eligible free credits for the selected video settings. Free credits are not unlimited videos: grants are restricted by region, sign-in method and anti-abuse checks, and not every account receives a grant. Email-and-password registration does not receive the registration gift. Check your available balance and the model’s credit estimate; a free balance can be too small for a video. The model overview on this page lists costs from the current settings.

Can Nano Banana make videos?

No. Nano Banana, including Nano Banana Pro and Nano Banana 2, is Google’s image model: it creates and edits still images. To get a video, make or edit your image with Nano Banana in the AI image generator, then open the video generator, choose Frames, upload that image as the first frame and animate it with Seedance 2, Veo 3, Kling 3.0 or MiniMax H3. Each of these video models can also start from a text prompt.

What is an AI video generator?

An AI video generator creates short clips from a text description or animates a reference image. Image to Image AI lets you choose a supported video model and review its available settings before generating.

Can I create video from text prompts?

Yes. Choose a video model that supports text-to-video, describe the scene, camera movement, subject, style, and duration, then submit.

Can I animate an image into a video?

Yes. Every model in the video generator accepts an image as the first frame. Choose Frames, upload your image, describe the motion you want, and keep the prompt focused on camera direction and subject movement rather than redesigning the image.

Where do generated videos appear?

Videos take a while to generate, so they render in the background after you submit. When a video is finished, open My Videos to review it.

How do I choose the right video model?

Select a model in the video generator and check its supported inputs, duration, resolution and credit estimate. Keep the same reference and motion prompt when trying another model so you can judge the difference. The separate model comparison tool compares images, not videos.