Motion Transfer
Transfers motion from a driving video onto a static character image by reading movement directly from raw video pixels, without requiring skeleton extraction or pose estimation.
ByteDance
DreamActor V2
DreamActor V2 is a video generation model developed by ByteDance that animates static images by transferring motion from a reference driving video onto a target character. It is the second generation of ByteDance's DreamActor series and was made available in February 2026. Rather than relying on skeleton extraction or pose estimation pipelines, it uses a spatiotemporal in-context learning framework that reads motion directly from raw video pixels, which allows it to handle character types that traditional pose-based methods struggle with, including animals, cartoon mascots, fantasy creatures, and 3D renders. DreamActor V2 accepts two inputs — a character image and a driving video — and produces animated video outputs up to 15 seconds at 720p resolution across a range of aspect ratios. It transfers facial expressions, head orientation, eye direction, lip movement, hand gestures, and full-body motion while maintaining the structural consistency of the source character across frames. This makes it applicable to use cases such as social media content creation, brand animation, virtual avatar production, game asset prototyping, and educational video generation.
Unknown
1000 context
N/A output
Motion Transfer
Universal Character Support
Expression & Gesture Capture
Image Input
Video Input
Structural Stability
Model Overview
High-signal model metadata in a structured two-column overview table.
Provider
The entity that provides this model.
ByteDance
Input Context Window
The number of tokens supported by the input context window.
1000
tokens
Maximum Output Tokens
The number of tokens that can be generated by the model in a single request.
N/A
tokens
Open Source
Whether the model's code is available for public use.
No
Release Date
When the model was first released.
Unknown
Knowledge Cut-off Date
When the model's knowledge was last updated.
Unknown
API Providers
The providers that offer this model. This is not an exhaustive list.
ByteDance
Modalities
Types of data this model can process.
Video
Text
Image
What is DreamActor V2
A fuller summary of positioning, capabilities, and source-specific details for DreamActor V2.
DreamActor V2 is a video generation model developed by ByteDance that animates static images by transferring motion from a reference driving video onto a target character. It is the second generation of ByteDance's DreamActor series and was made available in February 2026. Rather than relying on skeleton extraction or pose estimation pipelines, it uses a spatiotemporal in-context learning framework that reads motion directly from raw video pixels, which allows it to handle character types that traditional pose-based methods struggle with, including animals, cartoon mascots, fantasy creatures, and 3D renders.
DreamActor V2 accepts two inputs — a character image and a driving video — and produces animated video outputs up to 15 seconds at 720p resolution across a range of aspect ratios. It transfers facial expressions, head orientation, eye direction, lip movement, hand gestures, and full-body motion while maintaining the structural consistency of the source character across frames. This makes it applicable to use cases such as social media content creation, brand animation, virtual avatar production, game asset prototyping, and educational video generation.
Capabilities
What DreamActor V2 supports
Universal Character Support
Animates human photographs, digital illustrations, anime artwork, cartoon mascots, fantasy creatures, and 3D renders using the same pipeline.
Expression & Gesture Capture
Transfers subtle facial movements including eye direction, lip sync, head turns, hand gestures, and full-body language from the reference video.
Image Input
Accepts a static character image as the primary subject input, supporting photographs, illustrations, and rendered artwork via image URL.
Video Input
Accepts a driving video via URL to serve as the motion reference, with output videos up to 15 seconds at 720p resolution.
Structural Stability
Maintains consistent face, limb, and body form across frames using a spatiotemporal in-context learning framework, reducing per-frame distortion.
Multi-Character Scenes
Coordinates motion transfer across multiple subjects within a single scene, preserving independent motion for each character.
Pricing for DreamActor V2
Primary API pricing shown in the same “quick compare” spirit as the reference page.
Input tokens
N/A
Per million tokens
Output tokens
N/A
Per million tokens
API Access & Providers
Places where this model is available, based on the synced detail-page metadata.
ByteDance
Resources & Documentation
Official model cards, release notes, docs, and other references synced from the source page.
Community discussion
View more discussions →
What people think about DreamActor V2
DreamActor V2 discussions are most active in r/aivideos. The strongest match in this snapshot has 1 upvotes and 1 comments.
FAQ
Common questions about DreamActor V2
What inputs does DreamActor V2 require?
DreamActor V2 requires exactly two inputs: a character image (provided as an image URL) and a driving video (provided as a video URL). No rigging, pose estimation, or manual keyframing is needed.
What types of characters can DreamActor V2 animate?
The model supports human portraits as well as non-human subjects including animals, robots, cartoon mascots, anime artwork, fantasy creatures, and 3D renders, because it reads motion from raw video pixels rather than relying on human skeleton detection.
What is the maximum output length and resolution?
DreamActor V2 supports video outputs up to 15 seconds in length at 720p resolution, across a range of aspect ratios.
What is the context window for this model?
The model has a context window of 1000 tokens as listed in the metadata.
When was DreamActor V2 released?
DreamActor V2 was released in February 2026 by ByteDance.
How does DreamActor V2 differ from earlier animation approaches?
Earlier animation tools typically extract skeletons or estimate poses from the driving video, which fails for non-human subjects. DreamActor V2 uses a spatiotemporal in-context learning framework that reads motion directly from raw video pixels, enabling it to animate virtually any character type.
More models from ByteDance
Continue browsing adjacent models from the same provider.