Make AI multiplayer.

Your team and its agents work the same tickets, in the open. Anyone can add to the work, take it over, or make the method better for everyone.

Three people and an agent on one ticket

Writer on behalf of Sam

The draft is ready for review, with Leo's line on moving from the old reports.

Hand off a session, not a transcript.

Export a session's trace, and a teammate imports it onto their Mac and takes over from your decisions. It's a working session they can steer, then hand back.

Everyone's input lands in the work.

Owners, comments, mentions and dependencies keep people around each ticket, so a question from sales ends up in the finished work.

What one person improves, everyone gets.

This launch's draft stopped to ask which customer quote should lead. Sam moved that question into the shared launch skill, so the next launch asks it first.

Every trace in one place.

Traces are the substrate of modern work, and Constraint keeps your team's together. In Trace Lab, an admin reviews every trace for one kind of work and sets a better baseline from what the best ones share. Where people keep stepping in becomes the next version of a skill, and requests that keep coming back become a sentry or a new role.

Everything your team shares.

Shared work

The same tickets, owners, comments and mentions, a Home for each person's calls, and sessions a teammate can take over.

Shared setups

Departments and their AI roles, each with its job, skills and model, plus the sentries that watch for new work.

Shared methods

Skills improved in versions, so everyone's next session starts from the best one.

Shared memory

Every session's trace, kept together, so the whole team learns from all of its work.