AI AvatarMotion API

Overview

AvatarMotion Multi works in two steps. An image model places the subjects from your photos together in one scene, then a video model turns that still into a short clip. Which models run depends on the preset. Polaroid and Ride an Animal pair Seedream 5.0 Lite with a Wan image-to-video model, and the six pet themes pair Seedream 5.0 Lite with Kling 2.6. The family, pet pose and Speed Drifter presets pair Nano Banana Pro with Seedance Pro. The three Drama presets skip the image step and go straight to Seedance 2.0. You choose a preset and upload the photos it calls for.

What you can build

The 30 presets:

• Polaroid booth (9): two people inside an instant-photo frame, smiling, hugging, kissing, shaking hands, high-fiving, adjusting hair, making a heart, showing a victory sign or giving a thumbs up.
• Pets (11): people with their pets, either in themed settings (Ancient Rome, ocean, frozen, jungle, desert, volcanic) or in poses like Happy Paw Lift and Nose Touch Bond.
• Family photoshoot (5): couple and group poses such as Tender Forehead Kiss and the three-person Joyful Family Pile.
• Drama (3): two people in an apartment, under water or caught in a tsunami, with sound. Only the first photo you upload reaches the video model, and the second is not used.
• Ride an Animal: one sample shows a woman sitting on a lion.
• Speed Drifter: the only preset that starts from a single photo, of a car.

Inputs

• Photos: one to three, depending on the preset, in JPG, PNG, WebP, GIF or HEIC (the upload field holds up to four). Most presets want two people. The pet themes want one person and one pet, Double Kiss Pet and Pet Triangle Portrait want two people and a pet, and Wink Pet Stack wants one person and two pets.
• Effect Type: the preset.
• Seed, under the advanced settings: a number from 0 to 9999999, default 42.

Outputs

A short video, plus the still image it was animated from for every preset except Drama. Polaroid and Ride an Animal clips render at 480p. The family, pet pose and Speed Drifter presets return five seconds at 480p, with family and pet pose clips in portrait. Drama returns five seconds at 720p and the pet themes five seconds, both with sound. The four samples on this page (three Polaroid scenes and one Ride an Animal) are larger than what these presets return today: about five and a half seconds at 720 pixels tall, from square to wide.

Safety & compliance

Only upload photos of real people when you have their consent.