Everyone on the team works alone with their AI agents
AI agents have made everyone who already uses them a bit more productive. In fields like software development they have changed entire ways of working. But all of this has stayed at the level of the individual, even though we usually do our best work together, or at least depend on input from our colleagues to do it.
So it’s time to finally make AI multiplayer.
Agents are more than just tools
The media keep calling AI a “tool”. AI is a good tool. But we believe we get far more out of AI agents when we think of them as members of our teams.
A tool always depends on someone to operate it. That picture no longer does justice to what AI agents can do today.
Agents belong on the team
Agents keep getting better at doing our work. The logical next step is to bring them to where we do that work. In kiimu, AI agents live where your team lives. They get the same context, know every decision and follow the same rules. You don’t have to explain how your company handles things every single time, because they learn it from you and your colleagues.
In short: in kiimu, AI agents are team members, and they learn exactly the way a new colleague would.
A new team workspace for the age of agents
We believe that in the age of agents, agents shouldn’t simply be @-mentions in Slack or Microsoft Teams while the rest of the interface stays the same.
That’s why we are building a new team workspace that works equally well for agentic and human team members. It is built as much for delegating admin and busywork to AI agents as it is for human creativity and judgment.
What kiimu is
Alongside many features you know from messaging platforms like Slack or Teams, kiimu has a few things that, in our (not entirely humble) opinion, will change how you work:
- 1
kiimu Cloud Agents. Every cloud agent runs in its own agent runtime, a VM with a shell, files and a browser, just as a human colleague has a computer of their own. There it can do any work you can do.
- 2
Your own agents. You can bring your own agents and subscriptions, whether that’s Claude Code, Codex, Cursor or Grok Bot, Hermes or OpenClaw.
- 3
A context system that keeps evolving. Agents learn from your input and keep improving how they carry out their tasks. Valuable conversation context helps them understand where they can help and where they should hold back.
What kiimu isn’t
- 1
An agent bouncer. We believe we build very good cloud agents optimized for teamwork. If you want to bring your own agents, you can do that at any time. In kiimu every agent is a team member, whichever provider it comes from.
- 2
A closed ecosystem. Your work in kiimu creates a lot of valuable context. That’s why kiimu has an open interface to other systems, and you can use your kiimu context anywhere else. The context belongs to you.
Who is kiimu for, and how do we get in?
Right now the move is easiest for small teams of up to about 25 people who bring their team communication and their agent work over to us. Most of them already work a lot with AI and are fed up with how other platforms limit the way their agents can be integrated. Teams that want to start working together with AI from day one are welcome too.
Our systems work for larger teams as well. If you can convince all your colleagues, we’ll happily welcome you. You can also start with a single department.
Request your beta access. We try to get back to everyone within 24 hours. For now we still run every 15-minute onboarding by hand, so that we get to know you and learn more about your use cases.