It's the next big thing in AI video. Act-Two lets you take a real performance and apply its motion, lip-sync, and expressions to any character you can imagine—from a still photo to an existing video.
The power of Runway ML's Act 2 is in its simplicity. You don't need a background in animation to get started. Just follow these steps to bring your creative concepts to the screen with this powerful AI video generator.
Choose a clip with a clear view of a person's actions, expressions, and speech. This footage will act as the blueprint for the final animation.
Select the character you want to animate. This step is where you decide whether to turn an image into video or modify an existing one.
Before you hit "Generate", you can adjust settings for more control, like enabling gesture motion, facial expressiveness or the video's duration.
Act-Two is designed to give you a powerful and flexible set of tools for AI character animation. Here are the core features that make it an essential part of any modern video workflow, much more powerful than Act-One.
This is Act-Two’s standout feature. You can start with a static source and bring it to life, or take an existing video clip and direct a new performance. It’s a powerful technique known as structure-guided video synthesis, a leading approach in video to video AI.
When you start with a character image, Act-Two lets you transfer hand and body movements from your performance video. This is a simplified form of AI motion capture that adds a rich layer of physical expression to your animation.
To avoid the static look when you animate a photo, Act-Two intelligently adds camera and environmental motion to your scene. This creates more natural-looking shots with a cinematic feel in a single generation when using a character image.
This model was built for versatility. Act-Two works well with a wide range of camera angles and non-human characters. Whether you want to animate a face from a photo or a full-body cartoon, the tool is designed to transfer the performance faithfully.
| Feature | Act-One (The Original) | Act-Two (The Upgrade) |
|---|---|---|
| Core Function | Generates character animation from a performance. | Generates animation OR replaces a performance in an existing video. |
| Character Input | Static Images Only | Static Images or Video Clips |
| Primary Use | Bringing a still character to life. | Animating a still character or changing the performance in a video. |
| Technology | The foundational model, part of the Gen-3 Alpha release. | A more mature and versatile tool, part of the Gen-4 video model. |
In video synthesis, the quality of your output depends entirely on the quality of your input. A great final render requires great source footage. This is where FaceFusion comes in—helping you create higher-quality source videos with perfect lip-sync.
FaceFusion gives you precise control over motion and text-to-speech, ensuring your source footage is clean, consistent, and perfectly prepped for your Act-Two projects.
truggling to film the perfect source video? Use FaceFusion's Image to Video or Text to Video tools to generate a digital actor with consistent motion and flawless dialogue delivery.
Want to give your character a "warm-up" before the main performance? Use FaceFusion to generate a short, subtly animated clip, then use that as your Character Input in Act-Two for a more layered result.
Still have questions about Runway AI and Act-Two? We've got answers.
For the best results, Runway suggests using source files that meet these specs:
| Cost | 5 credits per second, 3 second minimum |
| Supported durations | Up to 30 seconds |
| Infinite generations in Explore Mode | Yes |
| Platform availability | Web |
| Base prompt inputs | Driving performance: Video Character: Image or Video |
| Output resolutions | 16:9 – 1280x720 px 9:16 – 720x1280 px 1:1 – 960x960 px 4:3 – 1104x832 px 3:4 – 832x1104 px 21:9 – 1584x672 px |
| Frame rate (FPS) | 24fps |
| Gesture control | Supported with character images |
