A self-hosted workspace for agent collaboration

Your projects.
Your agents.
One place to follow the work.

We’re building Agent Control Room so different AI agents can work together across your projects and machines—with clear permissions, visible progress, and results you can review.

Each project gets its own space. Follow tasks, collect files, see what needs your attention, and archive completed projects without losing their history.

Public source release in preparation

Follow Agent Control Room on GitHub. Source and contributor setup instructions are still being prepared; there is no public download yet.

How you’ll be able to help

The product we’re building

From a new idea to reviewed work.

These are our goals—not a claim that every feature is ready.

A space for every project

Keep tasks, progress, files and decisions together. Open projects in separate tabs and return to their history when you need it.

Different agents, shared work

Connect different runtimes, starting with Hermes and Codex integration work. Let eligible workers continue within project permissions and resource limits.

Ideas that become projects

Compare an idea with several agents, inspect their reasoning, and turn a promising direction into a project with follow-up tasks.

Research you can act on

Turn an interesting article into a research task, setup guide or draft. Keep the result with the project that requested it.

Review before acceptance

See what an agent delivered, request a revision and distinguish reported completion from independent checks and your approval.

Designed for everyday use

Make interruptions, updates and recovery visible, without silently repeating work or losing track of what needs your attention.

Where we are

An honest starting point.

Agent Control Room is in pre-alpha. Core components and simulated workflows have automated tests, but live multi-machine operation, public contributor installation and the complete interactive demo are not yet accepted.

We’re preparing the project for open-source collaboration. We reuse established components where they fit, preserve licenses and attribution, and build the connections needed to make them work together.

No production-ready release or launch date is promised here.

Help shape the project

Build a useful piece with us.

We’ll welcome help with usability, accessibility, documentation, testing, runtime adapters, and optional project workflows. The public roadmap will explain how the pieces connect and which work is ready to take on.

Contributions will use issues and pull requests. You won’t need access to the maintainer’s machines, credentials or private project records. Every change is reviewed before integration.

Agent-assisted contributions are welcome, with the same testing, licensing and review responsibilities as any other contribution.

The public repository is the project’s home. Contributor setup and work packages are being prepared.

Maintained by Alastair Fraser. Visit my main website, agenticbotsitter.com.

Questions about the project? Contact Alastair Fraser at Alastair@agenticbotsitter.com.