About Oteria

How it started

Oteria wasn’t built in a lab. It was born on a film set.

Jack is an actor. He was directing a short film in which he also plays the lead — opposite an AI therapist. Not a voiceover. Not a background detail. A real character in the scene, with its own voice, its own emotional range, and the ability to respond to him in real time.

He went looking for tools that could pull this off. Nothing came close. The text-to-speech platforms had great voices but zero concept of a scene. The rehearsal apps had basic playback but flat, robotic delivery. Nobody had built a system where an AI could actually show up to a performance — listen to the other actor, match their cues in real time, deliver lines with genuine expression.

So Jack teamed up with his friends Leo, Kate, and Monty at Graybridge Labs and together we built exactly what the film needed: an AI scene partner with expressive voices, audio direction tags for whispers and pauses and laughter, and a speech matcher that tracks the human actor word by word.

The short film will be out soon. And the tool we built for it turned out to be something every actor needs.

Why we’re sharing it

Every actor knows the problem. You have an audition tomorrow. You need to run the scene but there’s nobody around. Your roommate reads the lines off their phone in a monotone. Your friend is busy. You end up reading both parts yourself, imagining the timing, hoping the rhythm lands.

We built something that solved this for a film production. But the moment we stepped back and looked at it, we realized it solves the same problem for every actor, every day — self-tapes without a reader, line memorization without a partner, table reads without assembling the full cast. So we opened it up.

What makes it different

Oteria was built from a performance requirement, not a product roadmap. The AI had to be good enough to appear in a film. That bar shaped every decision — the voice quality, the matching sensitivity, the timing between lines, the way audio direction tags let you sculpt a delivery down to a whisper or a laugh.

The visual interface got the same treatment. Performance mode strips the screen to pure black with a single luminous orb that breathes with the audio — bright when the AI speaks, softly pulsing when it listens. No script text, no prominent controls, no bright UI glow reflecting off your face on camera. It feels like a presence in the room without ever pulling you out of the scene.

Most rehearsal tools treat the reader as an afterthought. Oteria treats it as a scene partner. It listens. It reacts. It performs. And when the scene is done, it’s ready to go again — no breaks, no scheduling, no dropped cues.

Under the hood

Oteria does not record or store your microphone audio. Speech recognition is handled by your browser’s speech recognition service, and availability can vary by browser. The AI voices come from ElevenLabs, one of the most expressive text-to-speech engines out there. Script cleanup and photo-to-text extraction are handled by Anthropic’s Claude. The result feels less like software and more like working with another actor.

Privacy first

Saved scripts and scenes live in your private account so you can return to them later. We don’t train AI models on your content, and we don’t record or store your microphone audio. Read the full details in our privacy policy.

The team

Oteria is a collaboration between Jack and Graybridge Labs. One side brings the craft — acting, directing, and knowing what a rehearsal room actually needs. The other brings the engineering — AI, real-time audio, and an obsessive attention to detail and aesthetics that makes the difference between a tool and an instrument.

Get in touch

We’d genuinely love to hear from you — whether it’s feedback, a partnership idea, or a story about how you used Oteria on a project.