Act-One, Explained Simply

Runway's Act-One is an AI tool that takes a video of a person talking or acting and maps their performance onto a new, AI-generated character. It is perfect for aspiring filmmakers, animators, or storytellers who want to create expressive, lifelike characters without needing a Hollywood budget or a team of 3D artists.

What it does well

Act-One is incredibly good at capturing real human emotions. Instead of just copying basic mouth movements, it tracks subtle facial expressions—like a slight smirk, raised eyebrows, or a genuine laugh—and transfers those exact feelings onto an AI character. This means the generated characters actually look like they are "acting" instead of just reading lines. It also handles complex camera angles and different lighting setups well, so the final video looks professional and cinematic.

How it compares

Compared to Runway's older video tools, which mostly let you generate short clips from text prompts, Act-One focuses entirely on performance. Instead of typing a paragraph and hoping the AI guesses what you mean, you now act it out yourself. This gives you way more control over exactly how the character moves and feels. It shifts the focus from "generating random video" to "directing a specific performance."

What it costs and what it can handle

While the exact price per video isn't detailed in the launch announcement, the tool is designed to take basic, low-quality webcam or smartphone footage and turn it into high-resolution, cinematic video. It can handle both close-up face shots and wider shots where the character is acting within a scene.

Worth knowing

Because this is a brand-new, highly advanced tool, it might take some practice to get your recordings just right. You will need to make sure your acting is clear so the AI can pick up on your expressions. Also, as with most new AI video tools, access might be limited at first while the company manages server demand, so you may have to wait in line to try it out.