AI Video Creator: Text or Photo to Video With Sound

The AI Video Creator makes short video clips with sound from a text description or a photo. It is powered by Google's Gemini Omni Flash. Describe the scene, the camera movement and the sounds, pick 4, 6, 8, 10 or 15 seconds in Wide 16:9 or Vertical 9:16, and choose Draft (360p), 720p or 1080p. Add up to 6 photos of people, pets or products to keep them consistent, then change the finished clip by describing the edit, make a vertical or wide version of it, or make it longer, up to 30 seconds.

How it works

Describe your video: the subject, the action, the camera and the sounds. Optionally add a start and end frame, or up to 6 photos under "Characters & objects", with up to 3 photos of the same person, pet or product under one name.
Video Creator makes the clip in a few minutes. It runs in the background, so you can keep browsing.
Watch, download or share the clip. Type a change into "Want to change something?", make a vertical or wide version, or tap "Make it longer" to add up to 10 seconds.

What people use it for

Short-form video for TikTok, Reels, and Shorts
Product and brand ad concepts
B-roll and atmosphere shots for editing projects
Storyboards and pitch visualizations that move

Frequently asked questions

What model generates the videos?

Google's Gemini Omni Flash makes every clip, with the picture and the sound generated together from your prompt and photos.

How long can a video be?

Each clip is 4, 6, 8, 10 or 15 seconds. A clip with an end frame, or a "Seamless loop" that starts from your photo, can be up to 10 seconds. "Make it longer" adds up to 10 seconds at a time to a finished clip, up to 30 seconds in total.

What does a 15-second clip cost?

A 15-second clip currently costs 500 credits at 720p, 700 at 1080p and 190 as a Draft (360p), more than a clip of 10 seconds or less. A 15-second clip can start from your photo and use photos under "Characters & objects". An end frame, or a "Seamless loop" from your photo, needs a clip of 10 seconds or less.

Can I change a video after it's made?

Yes. Type the change into "Want to change something?", such as "make it night" or "the dog wears a red collar", and Video Creator edits the finished clip. Changes work on clips up to 15 seconds long. You can add up to 3 photos to a change, such as your logo, and use their names in it, like "put Logo on the wall". It edits clips it made; it doesn't take a video you upload.

Can I get a vertical version of a wide video?

Yes. On a finished clip, "Make a vertical version" remakes a Wide 16:9 clip in Vertical 9:16 for phones, and "Make a wide version" does the reverse, at the same quality and length. It works on clips up to 15 seconds and costs the same as a change. The new version is a new render of the same scene, so check the details before you post it.

How long does a generation take?

A few minutes. It runs in the background, and the finished clip shows up in Recent on the tool page. Save it to your gallery to keep it; unsaved results stay available for 24 hours.

What makes a good video prompt?

Describe it like a director: the subject and action, the setting, one camera movement ("slow dolly in", "aerial shot"), the light, the mood and the sounds you want to hear. Put spoken lines in quotes.

Which quality should I pick, and what does a clip cost?

Use Draft (360p) to try a prompt, 720p for most feed content, and 1080p (upscaled from 720p) for bigger screens. Video Creator needs a Pro plan or higher; a clip currently costs from 350 credits at 720p and from 500 at 1080p, and failed clips are refunded automatically.

Can I use photos of people?

Use your own photos, or photos of people who agreed. Google blocks photos of some well-known people, and for users in the EEA, the UK and Switzerland it blocks photos of anyone under 18.