Shared documents changed how teams write. loop changes how teams work: one space where your whole team and its AI agents move the same work forward, in real time.
Tell the space what your team is working on. It carries that context across every run, for everyone.
Give it the skills the job needs, scoped to only the ones you grant.
Slack, Google Workspace, Microsoft 365. It acts in the real tools, not a copy.
Describe what should happen and when. You get a real, inspectable automation.
Every result is a durable file in a workspace you can export. Nothing is trapped.
Collaboration has always stopped at the document. Here, what you share is the work in progress: teammates appear live inside a running job, anyone can add a message to it, and speaker labels are preserved so it's clear who said what. When the work needs an answer, whoever is free gives it.
Collaboration that only happens while everyone is online is just a meeting. Work in loop runs in the cloud rather than in anyone's browser tab, so it continues overnight, over the weekend, and across time zones. The results are in the thread when the team comes back, even if every laptop was asleep.
Describe what should happen and when, and it becomes a real, scheduled routine your whole team can see and edit, not a script on one person's laptop. It opens in a visual canvas, and fires on a schedule, a webhook, a connector event, or a Slack or Teams @mention.
Your team keeps the tools it already has. Connect Slack, Google Workspace, and Microsoft 365, and loop acts inside them: it sends the real email, replies in the real thread, and edits the real document. On day one it reviews what you've connected and writes a briefing on your projects and stakeholders, so it's useful before anyone configures anything.
Describe the internal tool your team wishes it had and loop builds a real, installable app. No engineering backlog, no ticket queue. Publish it to your company catalog so anyone can install and run it, with review, permissions, and versioned changelogs built in.
The same space, shaped around how your team already works together. One moment from each.
Your data stays yours. loop runs on open-source models in an environment you control, so your work is never used to train anyone's model, not ours and not a third party's.
Run on open models you can inspect and self-host, with no black-box dependency on a single vendor.
Your content, files, and outputs are never used to train foundation models. Not now, not ever.
Data stays in your environment and region, isolated per workspace, with an immutable audit trail.
A new kind of online collaboration, currently in private preview. If you have an access code, come in. If not, leave your email and we'll open the door as we widen access.