ElevenLabs launches Actor Mode to direct AI voice acting

Telling an AI voice to sound sad or sarcastic using only text prompts has always been an exercise in frustration.

ElevenLabs just rolled out Actor Mode to everyone inside Studio, letting creators steer AI voice delivery using their own spoken audio.

Directing the AI with your own voice

Instead of endlessly tweaking text descriptors and regenerating clips, users can now speak the lines themselves to guide pacing, pitch, and emotional emphasis.

The AI keeps its selected target voice, but adopts the delivery and cadence from the human input. In the launch demo, ElevenLabs creator growth team member James used a Hamlet monologue to show how dramatic phrasing carries over into the generated audio.

Why it matters

Prompting audio through text has hit a clear ceiling when it comes to nuance. Words like dramatically or whispered rarely capture the exact micro-pauses or inflection a director wants.

We think speech-to-speech guidance is the most practical way forward for creative control in audio production. It gives filmmakers, game developers, and podcasters granular direction without forcing them to settle for random generations.

The catch is that the tool now relies directly on your own performance. If your guide recording sounds flat, the AI output will likely follow suit.