Join the waitlist

Work with your team and AI agents in one shared terminal.

Bring the agents you already use and invite your team. Everyone jumps into the same session, steers any agent, and builds on each other’s work at the same time.

Join the waitlist See demo Book a call

Early access for small teams.

You’re on a team. Working with AI should feel like it.

Today one person shares their screen and drives. Everyone else works with their own agent and pastes the results into the chat. The driver reads it all, sorts out what fits, and pastes it in. The rest of the team watches.

In the 2025 Stack Overflow survey, 69% of developers who use AI agents said the agents made them more productive. Only 17% said agents improved how their team works together. Source

Read the demo as text
  1. Friday, 5:04 pm. Refunds are failing in prod. ben asks his Claude why; the logs show refund errors since deploy 1.8.3.
  2. ben types invite in xmu, picks ana and mia from his team, and the session goes from private to shared.
  3. The invite shows up in ana’s and mia’s own terminals. ana joins with Codex, mia with Cursor.
  4. ana and mia join ben’s session, each with their own agent.
  5. ben’s Claude finds the cause: line 40 of limit.ts compares cents to dollars.
  6. ana types into ben’s pane and steers his Claude to fix partial refunds too.
  7. ana’s Codex writes the test while Claude fixes the code and opens a pull request.
  8. mia’s Cursor reviews the fix as it lands. mia: “Looks right. Ship it.”
  9. They merge the fix, deploy 1.8.4 goes out, and refunds work again.

In xmu, everyone drives.

Your teammates and their agents work in the same session with the same context. Agree on the approach, tell any agent what to do, and watch each result land where everyone can see it.

Pick up where your team left off.

The conversations, decisions and work so far stay in the session, for people and agents. Join late and you’re caught up without anyone explaining it again.

How it works

  1. Step 1Terminal

    Create a shared terminal.

    Run xmu in your repo. You get one terminal your whole team can join.

  2. Step 2Team

    Invite your teammates and their agents.

    Send a link. Each person brings the agent they already use: Claude, Codex, Cursor or any CLI agent.

  3. Step 3Together

    Start working together.

    Everyone types in the same session. Steer any agent, yours or a teammate’s, and build on each other’s work.

  4. Step 4Ship

    Ship together.

    Test, review and merge from the same place. The whole story stays in one history.

Works with your existing stack and any coding agent.

Your repo, your git, your terminal. Nothing to migrate.

Get early access to xmu.

We’re starting with a few small teams.

We’ll only email you about early access.