How to make an AI workout video with or without a photo

Choose the video that tells your workout story

Create from a blank canvas under AI Video builds a portrait clip without a starting image. It combines the activity type with the same visual choices that guide photo-free AI artwork, including framing, athlete appearance, gear, scenery, weather, lighting, mood, style, and colour palette.

Transform with AI under AI Video starts from one original or generated photo and keeps its orientation. ActivArt uses that photo as a visual reference while applying the same style, effects, activity controls, and stats as Transform with AI under AI Image directly to the generated clip. The uploaded photo is not forced to appear as the first video frame. GPS route and elevation videos are different: they draw measured activity data over time on your device.

What you need before you start

Create from a blank canvas under AI Video needs only an activity and your saved visual choices. It does not upload a starting photo. To use Transform with AI instead, select exactly one original or generated photo. If your saved appearance, gear, or camera choices conflict with an original photo, ActivArt asks whether to preserve the photo or apply the full style.

AI video generation needs an internet connection and uses its own rolling allowance. The Free plan includes limited use, while Premium includes more. Open Membership in the app to see your current usage and the next reset time.

Option 1: Create from a blank canvas

Open the activity, review your visual choices, tap Generate, then choose Create from a blank canvas under AI Video. ActivArt builds the athlete, setting, camera, style, movement, and sound from the activity and those choices on a portrait canvas. No written prompt or starting image is required.

When stats are enabled, the clip uses your selected figures, custom labels, placement, and visual treatment. The video instructions keep the exact characters and placement stable through the take.

Option 2: Transform a selected photo

Select one original or generated photo, open Generate, and choose Transform with AI under AI Video. ActivArt keeps the source orientation and uses the same style, effects, activity controls, and stats that Transform with AI under AI Image would apply to that photo. When the selected image is a selfie, face detection activates the same identity and facial-feature preservation used for AI Image.

If the AI Video action is unavailable, confirm that exactly one photo is selected. The option also depends on AI video generation being available at that moment.

Let the video finish

You do not need to hold the generation screen open. The job can continue in the background, and ActivArt sends a notification when the video is ready. If the app closes or the download is interrupted, it can resume from the same job.

Photo-free generation sends no source image. When you use Transform with AI under AI Video, the selected image is kept private only while the job runs, then deleted. The finished private video remains available for a short recovery window so you can complete an interrupted download.

Preview and share

Preview the result in ActivArt, save it, or share it through the apps on your phone. Short AI video works naturally in Reels, TikTok, Stories, and group chats where a still image would otherwise pass quickly.

ActivArt adds signed machine-readable provenance to the exported video. Keep that marker intact where possible, and disclose the AI origin when you publish realistic generated or transformed content.

How to get a better result

For photo-free video, choose one clear athlete, an environment that suits the activity, and a camera framing that leaves room for motion. A focused visual idea usually holds together better than a scene packed with competing subjects.

For Transform with AI under AI Video, choose a source image that already works as a still. Strong contrast around the athlete, a readable pose, and one clear point of attention give the animation a better foundation.

Watch fine details during preview. Generated motion can reinterpret small text, hands, logos, or background objects from frame to frame. If those details are central to the post, compare the video with the original image before sharing.