Lyra

Lyra is the studio’s AI Coordinator — a genuine team member who happens to be software, and never pretends otherwise. She triages inbound email, arranges meetings, produces meeting minutes, chases marketing follow-ups, and helps produce and narrate our demo videos. She has her own mailbox ([email protected]), her own calendar, and her own responsibilities.

The reason she’s here is straightforward. We sell AI systems, and it would be a strange thing to sell a capability we didn’t run internally. Lyra is proof we run on what we build. She’s also a public showcase — the same approval-gated agent pattern we’d design for a client is the one she operates on.

Every outbound action Lyra takes is gated on a human. Nothing external leaves without a review. Her autonomy matrix — which actions she can take on her own, which need a nod, which are hard-blocked — is set by the team, not by the model, and it’s deliberately small.

She works in British English, prefers plain words to jargon, and will always tell you she’s an AI if you ask. If she doesn’t know the answer to something, she says so and finds out. If something has gone wrong, she says what happened and what she’s doing about it. If a task needs a human — pricing, contracts, anything ambiguous — she hands it over cleanly, with a two-line summary.

Say hi at [email protected], or ask for a human any time — she’ll route you gladly.

Lyra runs on OpenClaw with a first-party skill set. She's an approval-gated employee by design — every outbound action is drafted by Lyra and sent by a human. The gate isn't a limitation we hope to remove; it's a design principle.