Grok Imagine — Image to Video

Animate a still image, up to 30 seconds.

6-30sNo Audio3-25 creditsImage to Video

Creative Vision

Describe how the image should move.

0/5000

Video Parameters

6s
6s30s

Longer videos cost more.

Fun exaggerates motion and expressions.

Production Cost
3
Credits Balance
Loading...
Model
Image to Video

Cinema Preview

Ready for Production
Configure your settings and begin creating your cinematic masterpiece

Generation History0 records

Latest 10 records • Available 7 days

Grok Imagine Image to Video Generator

Upload one still, describe the motion, and the Grok Imagine image to video generator animates it for 6 to 30 seconds at 480p or 720p. It is the cheapest way to animate a photo on this site: 3 credits for a 6-second clip, and nothing at all if the run fails.

Best for product shots, portraits, artwork and screenshots that need movement rather than a full re-shoot

Grok Imagine

What the Grok Imagine Image to Video Model Does

One image in, up to 30 seconds of video out. Your still becomes the opening frame, and the prompt decides what happens next.

Input

One Image, Any Subject

JPEG, PNG or WebP up to 10MB. Product photos, headshots, illustrations, screenshots and scans all work, because the model animates whatever is in the frame.

Control

You Direct the Motion

The prompt is required here on purpose. "He picks up the guitar and starts playing" gives a usable clip; leaving the box empty gives you drift and wobble.

Duration

6 to 30 Seconds

The other image-to-video models here stop at 10 or 15 seconds. This one runs to 30 in a single generation, which is enough for a full product turn or a held expression.

Resolution

480p or 720p

480p is the 3-credit draft tier for testing which motion prompt works. 720p is the one you publish once you have found it.

Framing

Framing Follows Your Image

There is no aspect ratio picker, and that is intentional. The upstream model ignores it when a single image is supplied, so the output keeps the shape of what you upload.

Credits

Failures Cost Nothing

Credits are held at submission and released automatically if the model errors out, returns no file or never answers. Only delivered videos are actually charged.

Animate a Photo for 3 Credits

Crop the image the way you want the video framed, write one clear sentence of motion, and generate. Iterate at 480p before spending on 720p.

One image, up to 30 seconds
Step by step

How to Animate an Image With Grok Imagine

Four steps, about a minute of waiting, and no editing software involved.

1
Step one

Upload the Still

Drop in one JPEG, PNG or WebP under 10MB. Crop it first if you want a different frame, since the video inherits the shape of the image you supply.

2
Step two

Describe the Motion

Say what moves and how: a head turning, fabric catching wind, the camera pushing in. Describe one clear action rather than three competing ones.

3
Step three

Choose Length and Resolution

Set the slider between 6 and 30 seconds, then pick 480p or 720p. The credit cost is shown live, so long 720p runs never surprise you.

4
Step four

Generate and Download

A 6-second 480p clip came back in roughly 30 seconds when we timed it. Longer clips at 720p take proportionally more, and results stay in your history.

Auto-playing timeline

Getting Believable Motion

The model respects physics it can see. A photo with obvious depth, a clear subject and room to move animates far better than a flat, busy composition.

One subject, one action
Grok Imagine

Grok Imagine Image to Video

FAQ

File limits, motion prompts, credit costs and framing rules for animating a still image with Grok Imagine.

It takes one still image plus a description of the motion and returns a video clip of 6 to 30 seconds. Your image becomes the starting frame and the model generates the frames that follow. FlowVeo3 is an independent platform and is not affiliated with xAI.

JPEG, PNG and WebP up to 10MB, one image per generation. Upload the highest-quality version you have, since compression artefacts in the source tend to be amplified once the frame starts moving.

Because it would do nothing. The upstream model ignores aspect ratio whenever a single image is supplied, and single-image input is all this endpoint accepts, so the output always follows the shape of your upload. Crop before uploading to change the framing.

Not on this model, which takes exactly one image. For first-frame and last-frame control, use Wan 2.7 Image to Video, which accepts both plus an optional driving audio track.

Yes. The upstream API treats it as optional, but we require it, because image-to-video runs without a motion description reliably produce unusable footage. One specific sentence is enough.

3 credits for 6 seconds at 480p, up to 25 credits for 30 seconds at 720p. Pricing is identical to Grok Imagine Text to Video, and new accounts start with a small free credit balance.

It is stored on our own object storage and handed to the model provider as a URL so the generation can run. It never appears in a public gallery and we do not use it for training. The storage path is served over a CDN, so treat the file link as unlisted rather than secret.

Animate Your First Image

One photo, one sentence of motion, 3 credits at 480p. Move up to 720p and 30 seconds once the prompt behaves.