Why we built Crilo

Design and planning tools were built for a world where every collaborator is a human clicking around a screen. That world is changing. Your most prolific teammate might now be an AI agent — and the agent has no cursor, no seat, and no patience for a tool it has to be taught to click.

Crilo starts from the opposite end. The canvas is a shared surface that a human watches and an agent writes to, at the same time, over a live connection. You describe a screen; the agent lays it out with real layout, type, and color. You move a card; the agent sees it. No handoff, no export, no “let me regenerate that.”

One canvas, two kinds of collaborator

• Humans get a fast, legible view of the work — boards, elements, tasks — that updates the instant anything changes.
• Agents get a small, sharp set of tools over a remote MCP endpoint: create elements, lay out groups, draw connectors, track tasks.

The result is a work surface where planning, design, and tracking live in the same place — and where the agent you already use is a first-class participant, not a bystander.

No login, just a link

A workspace is an unguessable link. Share it and someone’s in — no accounts, no invites, no friction. When you need real access control, turn on login and the same canvas gains members and grants. Start open; lock down when it matters.