Image to video turns one still into a short clip.
OpenUpload one photo, choose a video model, and generate. The button shows the credits for that resolution and length.
Audio is included. This model does not turn it off.
Write a prompt.
The source is a single photograph, not a storyboard and not a line of text. You write how that picture should move. The person or product in the still is the subject of the clip. Length, frame, and quality are the options that model lists, and the button adds them into one credit count.
Frame, length, and quality are the options that model lists. The page has a model menu because each video model has its own price. A still edit stays on image to image.
Before the run, the button reads the credit cost. If the account does not have enough credits, the path is a plan or a pack, not a half-finished video. Monthly credits last a month. Yearly credits last a year. A pack does not expire and is not refilled every day. A new account starts with 4 credits.
If the video fails, the credits for that run return without a request. A new run spends its own credits. The source photo is removed from storage within an hour after the run finishes. It is not used to train photoeditor models.
The clip starts from the photo you upload.
The sentence is the motion. The photo stays the subject.
Only durations that model can bill appear on the controls.
Download it, or leave it on the account for the plan’s retention.
Image to video starts from one still and returns a short clip. You pick a video model, then the frame, length, and quality that model lists. The button shows the credits for that combination before the run. This page does not offer a length or a resolution that has no catalog price.
The still and the video stay together after the run. You can download the clip from the stage or find it later in Creations. A failed video returns its credits on its own. The upload is deleted within an hour after the run finishes. This page does not make a video from text alone. Grok Imagine and Veo 3.1 take one still. Grok Imagine 1.5 can take up to three.
Length and resolution change the provider cost, so they change the credits. The button follows the model and the settings on screen. It does not stay on one number while you change the length.
The still you upload is the subject. A clip from words alone is the text to video page, and that button shows its own credit cost. A still edit before the motion stays on image to image.
Upload one JPEG, PNG, or WEBP. Write the motion. Choose a model, then the frame, length, and quality it lists. Read the credits on the button. Generate. The stage keeps the still with the clip.
One photo is the first frame.
Write how the frame should move.
The button adds those into the credit count.
A failed clip returns the credits it spent.
The useful case is a product, a portrait, or a room you have already accepted as a photograph. The clip is a few seconds of motion on that frame. Grok Imagine and Veo 3.1 take one still. Grok Imagine 1.5 can take up to three.
Commercial use follows the still. If you have the rights to the photo, you can use the clip. The page does not grant rights to a face or a product you did not own. The credits come out of a monthly plan, a yearly plan, or a credit pack.
Lite $17 / Plus $33 / Pro $72 — full list on pricing
15 seconds, on Grok Imagine and Grok Imagine 1.5. Grok Imagine lists 6, 10, and 15 seconds. Grok Imagine 1.5 also lists 8. Veo 3.1 Lite, Fast, and Quality list 4, 6, and 8 seconds.
Grok Imagine 1.5, up to three. Grok Imagine and each Veo 3.1 tier take one still.
No. Veo 3.1 stays one price across 4, 6, and 8 seconds at 1080p. Grok Imagine is priced by the second at 480p or 720p, so a longer clip does change that button.
No. This page needs a still. A clip that starts from a sentence is text to video.