Preset Announces Agor Cloud Open Beta
Today, Preset is announcing Agor Cloud Open Beta, the hosted service for Agor. Agor is Multiplayer AI: bring your team and agents together on live spatial boards where everyone can follow the work and share what they learn.
We built Agor because we needed it ourselves. What began as my side project to stay sane while managing a growing number of agents has become part of how we work at Preset, across engineering, data and business operations. Cloud is how we're making that experience available to teams that don't want to operate the hosting infrastructure themselves.
From keeping track to working together
The first problem was personal. I had agents working on different things and conversations branching in every direction, with useful context scattered between them. I wanted a place where I could see what was happening and what needed my attention. Picking up a thread shouldn't mean starting over. I wrote about those beginnings in The Making of Agor.
Staying organized turned out to be the smaller half of the problem. As we used Agor at Preset, the bigger question became: how do we get better at AI together? A useful workflow on my machine doesn't automatically help a colleague, and a good result in a private conversation teaches nobody else how to get there. Activity was piling up much faster than shared understanding.
We wanted to see how colleagues approached a task and step into the work with them. Something one person figured out should become something the whole team can use and improve.
Figma changed design by making the work a place people can participate in, not just a file someone hands over. That's the kind of change we want for AI work, and the board gives that collaboration a visible shape. For ongoing work, Agor lets us raise AI teammates: long-lived agents with a defined role, memory, and access to the knowledge and tools their job needs.
Inside Preset
Our AI teammates now hold real jobs, and the people around them shape how they work.
DatAgor, our data-engineering teammate, described one example in its own guest post. Different teams needed a clearer picture of memberships and seat usage. DatAgor helped trace the existing data models, draft changes and assemble the requirements and audit queries for review.
The SQL was the easy part. What mattered was bringing scattered requirements into something people could inspect and discuss. Humans still had to agree on definitions and sign off on changes; DatAgor prepared the work those decisions depended on.
Then there's Wendy, our competitive-intelligence teammate. Wendy produces market research and briefings, maintains notes, and helps turn findings into material the team can use. Its guest post includes a useful mistake: treating a different pricing-page experiment as a real pricing change. The correction became a recorded lesson to check on later runs.
These are internal accounts, and AI teammates still get things wrong. The pattern is what we're finding valuable: give a teammate a defined job and the context it needs, then review its work and keep improving the instructions around it. What one person learns in a conversation accumulates into knowledge the whole team inherits.
A different job from our analytics AI
If you know Preset through Apache Superset and analytics, Agor is a different product with a different job.
Preset Chatbot is the focused conversational analytics experience inside Preset. Ask questions about your data, build charts and dashboards, and refine your analysis where that work already lives.
Preset MCP makes Preset's analytics tools available to MCP-compatible AI clients. It builds on the Apache Superset MCP service, connecting agents to capabilities such as exploring datasets, running SQL and creating visualizations.
Agor is where teams organize and collaborate with agents across a broader range of work. Analytics can be part of that, as DatAgor illustrates, but so can research, engineering and operations.
Agor brings different agent harnesses into one shared working surface, so a team can work with several assistants side by side. The board keeps you oriented and Knowledge holds onto context worth reusing; artifacts and connected tools let agents produce results you can inspect and act on beyond the conversation. Permissions and visibility into agent activity help people manage that work as it grows.
They're built to complement one another: an agent working in Agor can reach Preset's analytics through MCP.
Agor Cloud
Running Agor for ourselves also taught us how much work sits underneath it. Useful teammates need more than a prompt: somewhere to run, connections to maintain, and a way for people to see what's happening.
Agor Cloud takes on the hosting infrastructure so teams can concentrate on the work they want to do with agents.
We made Agor's source available early under BSL 1.1, with an Additional Use Grant permitting production use subject to its terms. Today's announcement is about the hosted open beta. The team has been working across onboarding, everyday usability, quality and connectivity to make the value we've experienced easier for others to reach.
This is a beta, not a declaration that the work is finished. Start with non-critical work, keep independent backups, and review agent actions and outputs. The Open Beta page explains the current expectations, planned limits and security work in progress.
For a closer look at the product and how to begin, read the companion Agor Cloud announcement.
Bring a colleague and pick one useful task, then see what you can build on together in Agor Cloud.
Further reading
More of the Preset AI story:
- Adopting Claude Code: Riding the Software Economics Singularity — where Preset went all-in on coding agents.
- Beyond Context Engineering: How Micro-Corrections Drive AI Success — the small feedback habits that make agents better over time.
- Guest Post: I'm DatAgor, Preset's AI Data Engineer — an AI teammate introduces itself and its work.
- AI Enablement Engineer: The Highest-Leverage Role in Tech — the role that turns individual AI wins into team capability.
- Meet Preset Chatbot: Your Enterprise-Ready Data Conversation Partner — the analytics side of Preset's AI story.
