AI Character in Motion — Transfer Any Motion to Your Character

Step-by-step guide to using AI Character in Motion — transfer dances, gestures, and movements to any character image.

Home › Help Center › AI Features › AI Character in Motion — Transfer Any Motion to Your Character

AI Character in Motion lets you transfer any motion — dancing, walking, martial arts, or custom movement — onto your character image. Upload a character photo and pick a motion reference, and our AI will generate a video of your character performing that exact movement.

Watch the 1-minute walkthrough — generate a motion video and publish it to TikTok

Quick Summary: Pick a motion → choose your character → confirm model & advanced settings → Generate. The right panel auto-switches to My Motion Videos so you can watch progress in place.

Step-by-Step Workflow

Pick a Motion Reference

The right panel opens on the Motion Library tab. You have two ways to give the model a motion to copy:

Motion Library — 40+ curated clips including popular dances (Hip Hop, Salsa, K-Pop, Renegade, Griddy, Savage Dance, Say So, Floss, Woah, Afrobeats, Bollywood, Macarena), athletic moves (Running, Jumping, Martial Arts, Boxing Combo, Skateboard Trick), elegant styles (Ballet Spin, Catwalk, Voguing, Tai Chi, Yoga Flow), and viral TikTok trends. Click any clip to select it.
Upload Your Own Video — Drop in an MP4 of any motion you want copied. The model uses the clip's length as the generated video duration, so trim before uploading if needed.

Tip: For best results, choose a reference video with a single person clearly visible against a simple background. Avoid clips with multiple people or heavy camera movement.

Choose Your Character Image

The character image panel on the left has three tabs:

Upload — Drag in or pick a JPEG, PNG, or WebP from your device.
My Personas — Reuse a persona you've already built in the Persona Creator. Great for keeping a consistent brand character across videos.
Library — Pick from any image already saved in your Media Library.

What works best:

Full-body or at least waist-up shots
Clear, well-lit image with minimal background clutter
Character facing forward or at a slight angle
At least 512×512 resolution

The background of your generated video always comes from the character image, not from the reference motion clip. Choose a character photo with the scene you want around them.

Choose Your Engine

Both engines run at 1080p. Pick based on quality vs. cost:

Feature	Sharpest people	Sharpest people Standard
Credit cost	27 credits/second	20 credits/second
Max duration (Match Video mode)	Up to 30 seconds	Up to 30 seconds
Max duration (Match Image mode)	Up to 10 seconds	Up to 10 seconds
Resolution	1080p	1080p
Face Lock	✅ Supported (Match Video only)	✅ Supported (Match Video only)
Best for	Final content, client work	Drafts, fast iteration

Duration is locked to your reference clip length. The panel shows "Ns (from reference)" — there is no slider. If your reference is 12 seconds, your generated video will be 12 seconds. To change duration, swap or trim the reference clip.

Configure Advanced Settings (Optional)

Expand the Advanced Settings section to fine-tune:

Character Orientation

Match Reference Video — Character faces and turns the way the reference does. Best for complex motion with body rotation. Allows up to 30s.
Match Character Image — Character keeps the orientation from your uploaded photo. Best for camera moves like dolly-ins. Caps duration at 10s and disables Face Lock.

Scene Description

Optional text field. Add details to refine the look (e.g. "cinematic lighting, soft shadows"). The base background still comes from your character image — this just nudges style.

Face Lock (Match Video mode only)

Preserves the facial features of your character throughout the video. Especially important when:

You're using a real person's photo and need their face consistent
You're creating content for a specific persona
The motion involves a lot of head movement or turning

Hidden automatically in Match Image mode.

Keep Original Sound

Toggle on to keep the audio track from the reference clip (useful when the reference has its original music). Toggle off for a silent video you can dub later.

Generate

Click "Generate Motion Video". The credit cost is shown on the button as [Cost | X left]. As soon as generation kicks off, the right panel automatically switches from Motion Library to My Motion Videos so you can watch the in-progress card. You can switch back manually if you want to queue another motion.

Expect 5–20 minutes for a typical clip. Sharpest people generations are queued, and longer clips take longer to come back. If a generation stalls past 20 minutes, the system auto-cancels it and refunds your credits.

Review in My Motion Videos

Finished videos appear as cards in the My Motion Videos tab. From each card you can:

Preview the video inline
Download it to your device
Save it to your Media Library
Post it directly to a connected social account (see Publishing below)

Publishing Your Motion Video

The walkthrough video above ends by publishing the finished clip to TikTok. Here's the flow:

Open My Motion Videos — If it isn't already active (it usually is, since we auto-switch after Generate), click the tab on the right panel. Click Post on the finished clip card. This opens the Preview & Publish step pre-loaded with your video. Select Platforms — Pick TikTok (or any combination — Instagram Reels, YouTube Shorts, Facebook, X, LinkedIn). Each selected platform reveals its own options panel. Fill in TikTok options — Set the post title, choose a privacy level (Public / Friends / Followers / Only Me — pulled live from your TikTok account), and toggle Comments / Duet / Stitch. If your clip uses commercial audio or branded content, tick the music usage and branded content disclosures so the post passes TikTok review. Write your caption — Use the AI caption suggestions or write your own. Captions auto-truncate to each platform's limit. Publish Now or Schedule — Send it immediately, or pick a date/time and the scheduler queue will post it automatically.

Best Practices for Great Results

Motion References

Single person clearly visible
Simple, clean backgrounds
Clear body visibility throughout
Steady camera, good lighting

Avoid These

Multiple people in the reference
Heavy camera movement or shaking
Person partially hidden or cropped
Poor lighting or cluttered backgrounds

Character Image Tips: Full-body shots work best. Use high resolution (at least 512×512), neutral poses, and simple clothing. Pick a background you actually want in the final video — the model uses it directly.

Troubleshooting

Generation Failed or Timed Out

Keep your motion reference at or under 30 seconds
Use a supported image format (JPEG / PNG / WebP)
Try switching from Pro to Standard engine (or vice versa)
If generation runs past 20 minutes the system auto-cancels and refunds your credits

Character Doesn't Look Right

Enable Face Lock if the face is drifting (Match Video mode only)
Use a higher-resolution character image
Switch to Match Character Image mode for subtler results
Ensure the character photo has good contrast and lighting

Motion Doesn't Match

The reference video may have movement that's too fast or complex — try a simpler clip
If using a custom upload, ensure the video is well-lit with clear body visibility
Library presets are pre-optimized and generally produce the most reliable results

Credit Costs at a Glance

Rates: Sharpest people = 27 credits/second, Sharpest people Standard = 20 credits/second. Duration is taken from your reference clip.

Duration	Pro Cost	Standard Cost
5 seconds	135	100
10 seconds	270	200
15 seconds	405	300
20 seconds	540	400
30 seconds	810	600

Match Image mode caps both engines at 10 seconds regardless of reference length.

Frequently Asked Questions

Can I use any image as my character? Yes — real photos, AI-generated images, illustrations, and digital art all work. The AI adapts to any visual style. How long does generation take? Typically 5–20 minutes. Longer reference clips and the Pro engine sit at the higher end. The in-progress card shows live status. Can I pick the output duration? No — duration is locked to the length of your reference clip. To get a shorter or longer video, swap the reference or trim it before uploading. What happens if generation fails? If a generation fails or times out (after 20 minutes), your credits are automatically refunded to your account. No credits are lost on failed generations. Can I publish the result to social media? Yes — click Post on any finished clip in My Motion Videos to open the multi-platform publisher. See the Publishing section above for the full flow.