Runway adds Voices to Act-Two

Runway just dropped a new feature called Voices for its Act-Two tool, aiming to fix one of the most awkward parts of AI character performance.

With Voices, creators can now tweak and change how their performance sounds directly inside the Act-Two workflow, making the voice actually fit the generated character. Runway also released a new Runway Academy guide demonstrating how to use it.

Matching the voice to the persona

When you use your own camera footage to drive an AI character, audio is usually an afterthought. You might look like an alien warlord or an animated mascot, but you still sound like someone sitting in a bedroom.

Bringing voice controls right into Act-Two lets creators skip the hassle of exporting audio tracks to third-party voice changers. You can now tweak the vocal delivery to match your character without leaving the video pipeline.

Why it matters

For solo filmmakers and creators, this is an immediate workflow booster. Directing both your visual performance and your character's voice in a single pass saves serious post-production headaches.

In our view, video generation is moving beyond raw visual fidelity into full-pipeline usability. An uncanny voice breaks character just as quickly as bad motion, so integrating audio directly into performance capture makes a lot of sense. That said, Runway has not detailed the exact range of vocal customization yet, so creators will need to test how natural these voice adjustments really sound.