Act-Two
Act-Two transfers a driving performance to a character image or video. The performance supplies speech, expression and movement. The choice of character input determines which motion controls are available.
Character images and videos give different control
With a character image, Act-Two can transfer hand and body gestures and adds environmental movement. With a character video, it keeps the source camera and scene motion while changing facial performance; gesture control is unavailable. A character video shorter than the performance loops forward and backward to fill the duration.Performance capture with Act-Two
Prepare the performance for the intended shot
Practical guidance. Choose a driving performance that matches the final scene’s timing. If the camera motion must stay fixed, check the video-input behavior before generating. Review lip movement, hands and the transition at any repeated segment before placing the result in an edited sequence.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API, Provider product |
| Model identifiers | act_two |
| Product duration | Up to 30 seconds at 24 fps.Performance capture with Act-Two |
| API price | $0.05 per second in the current model catalog.Runway Dev model catalog |
| Product billing minimum | Three seconds, at five credits per second.Performance capture with Act-Two |
Related models
A video-generation model that accepts text prompts or starting images.
Launch date not confirmedAPI · Provider productRunway’s avatar model behind Characters, supporting conversational personas and generated character videos.
Launch date not confirmedAPI · Provider product