AISVIT / AI Video / Image to Video

DreamActor M2.0 — Image to Video

Image to Video with DreamActor M2.0 in AISVIT. Animate still images into dynamic videos with AI. Add camera motion, subject movement, and cinematic transitions from a single source image.

About this model

Model for controlled character animation from a single image and a driving video, transferring facial expression, gestures, and body motion into a short clip.

When is this model useful?

DreamActor M2.0 works best when you already have a static character image and a separate reference clip that contains the movement or performance you want to transfer.

Best fit tasks

  • Bringing portraits, photos, illustrations, avatars, mascots, and cartoon characters to life.
  • Transferring facial expression, gestures, dance, walking motion, or acting performance from one video onto a different character.
  • Creating character-driven clips for social media, teasers, reactions, memes, and short performance-based scenes.
  • Animating animals or non-humanoid characters where standard pose-tracking workflows often break down.

Main advantages

  • It works not only with realistic humans, but also with drawings, cartoon characters, and animals.
  • It preserves character identity and visual style better than many motion-transfer workflows while copying the movement from the driving video.
  • It is useful for both close-up facial animation and wider full-body movement shots.
  • The workflow is simple: one image plus one video, without rigging, keyframe animation, or a heavy manual setup.

Limitations to know

  • This is not a text-to-video model: it does not invent the scene from scratch and requires both a character image and a driving video.
  • It works best when the main subject is clearly visible; crowded scenes, strong occlusions, or chaotic camera movement can reduce quality.
  • The input image needs a person or character, and the input video is limited to 30 seconds in this integration.
  • If camera angle, framing scale, or the number of subjects differ too much between the image and the driving video, the animation may feel less natural.

How to use this model

With DreamActor M2.0, you do not describe motion with text. Instead, you show the model how the character should move by uploading a separate driving video.

Simple workflow

  1. Upload the character image. The best starting frame is one where the subject is clearly visible, not accidentally cropped, and not overly blurry.
  2. Add the driving video, which is the source clip the model will use for facial expression, gestures, head movement, and body motion.
  3. If you do not want to keep the opening transition in the final result, leave Cut first second enabled.
  4. Check the driving-video duration before generating, because it affects both the expected result and the total cost.
  5. Generate the clip and, if motion looks noisy or unstable, try a shorter or cleaner reference video with clearer movement.

Supported inputs

  • Required: one character image in JPEG, JPG, or PNG format, up to 4.7 MB, with resolution from 480x480 to 1920x1080.
  • Required: one driving video in MP4, MOV, or WebM format, up to 30 seconds, with resolution from 200x200 to 2048x1440.
  • The cleanest results usually come from inputs where the character is well lit, not blocked by other objects, and clearly placed in the frame.
  • If the scene contains multiple subjects, it helps when the image and the driving video show a similar number of characters and a similar composition.

What you get

  • A generated MP4 video file.
  • Movement, facial expression, and performance transferred from the driving video onto your character.
  • A short character-animation clip suitable for social content, teasers, character showcases, or scene prototyping.
  • When Cut first second is enabled, a result with the opening transition removed if the model produced one.

Other workflows for this model

More Image to Video models

AISVIT pricing details

  • Fixed rate: 5 credits per second.
  • Shorter driving video = lower cost, longer driving video = higher cost.
  • Character type, image style, and the Cut first second option do not change the rate by themselves.