AI Video (Ai Live)

The AI Video tab holds templates that turn a photo into a video generated by the AI. You create a new AI template just like on AI Photo; the prompt editor is the same too. All video runs in the cloud and is paid.

Modes

The modes here are for video: Kling and Pixverse, in several versions.

Duration and quality

• Kling — quality 720 or 1080; duration 5 or 10 seconds.
• Pixverse — quality 360P / 540P / 720P / 1080P; duration 5 or 8 seconds.

Pixverse also offers additional effects and styles you can apply.

Deep Fake to fix faces

You can apply Deep Fake face restoration on video too — faces change a bit more on video, so it helps. Note that fix-faces costs extra, so turning it on recalculates the processing price.

Template preview

You can upload a template preview — a photo or video — shown here for the template.

One template only

Unlike AI Photo, AI Video supports only one processing per project — you can’t offer the guest several video templates; there’s just one.

The prompt

Edit the prompt the same way as on AI Photo. Here it matters to describe how the photo is animated — how the person moves, what they do, maybe a small scenario — keeping in mind it has to fit the short clip length. In the negative prompt, state what you don’t want to see in the final video.

Full video preset (edit the AI clip)

An AI video isn’t just the raw generated clip — it has its own video preset, a full editor like the Video VFX editor. On the AI video you can:

• add an overlay on top of the video,
• cut / trim the clip,
• add effects,
• and essentially everything you can do with a video.

The only things not available on an AI video preset are chromakey processing and 360 “little planet” processing.

How it runs

The frame is uploaded, queued, generated, and downloaded like other cloud AI. Expect it to take longer than a photo render — plan the guest wait and the render timeout accordingly. See Credits & errors.