Skip to content
About ConferenceOS

AI should do more than help plan the event.

ConferenceOS is built around a simple idea: your chosen AI should understand the live conference, complete governed work across it, and leave the organizer firmly in control.

Built for conference organizers by conference organizers.

We started our own conference, All Things AI, in 2025. Sixteen hundred people came. The next year it doubled — and everything we had built to run the first one buckled under the second.

20251,600 attendeesOur first year.
The next yearDoubledSame team. Same stitched-together tools.
  • Technical debt faster than we could pay it down

    Every integration became a manual reconciliation. Every question became a search across systems that disagreed with each other.

  • Our own money, held until it was over

    The ticketing platform sat on ticket revenue until after the event. The venue, the AV company and the caterer all needed paying well before that.

  • Someone else’s system, someone else’s data

    The operating record of our own conference lived inside products we did not control and could not fully take with us.

  • Nowhere to put the busywork

    Chasing speakers, reconciling registrations, copying between tools — all of it could only be absorbed by people, and the people were already at capacity.

So we built the system we wanted. One where the data and the deployment stay under the organizer’s control, the ticket money lands on the organizer’s own payout schedule, and the platform is agent-native from the foundation — so the mundane work can be handed to AI agents instead of to the team.

An event organizer stands at the back rail of a full session room, arms folded, watching her own conference run
The job we wanted back: standing at the back of the room instead of inside a spreadsheet.

A passion project with three decades behind it.

Mark Hinkle

Mark Hinkle

Founder, Peripety Labs · Co-founder, All Things AI · Former Executive Director, Node.js Foundation · Former VP of Marketing, The Linux Foundation · Former VP of Community, CloudStack (Cloud.com)

Mark has spent more than three decades building communities around emerging technology — through the early internet and open source, systems management and cloud computing, and now AI.

He co-founded All Things AI, which is where these problems came from, and founded Peripety Labs. ConferenceOS began as his passion project and still has that shape: built by the people running the event, for the people running the event.

The aim: the Savannah Bananas of AI events.

The Bananas took the fees, the fine print and the dead time out of baseball and made the whole thing fun. That is the target here.

Drop the overhead a ticket-management platform imposes and that money can go back into the event instead of being passed through as fees and service charges.

Fun is not a soft goal. An event that is enjoyable to run is a better event to attend — speakers get a team that is not underwater, sponsors get attention instead of apologies, and attendees get the conference the organizers actually meant to hold.

A packed conference session room seen from the front, the audience leaning in and paying close attention
The whole point of the software is this room.

Make the event operable, not merely searchable.

ConferenceOS gives AI the event-specific context and boundaries needed to be an operations partner, not another disconnected writing tool.

01

A conference is a system

Programming, people, money, partners, places, communication, and event-day decisions affect one another. The software should understand those relationships.

02

Organizer attention is scarce

AI should absorb reconciliation work and repetitive handoffs so the team can focus on judgment, experience, and the people in the room.

03

Each event should compound

Structure, workflows, decisions, and operating history should make the next conference better instead of disappearing across one-off tools.

Open at the edges. Governed at the core.

An agent-ready platform should preserve choice and control as models, providers, workflows, and conference formats change.

OPEN

Bring the AI you trust

Claude, ChatGPT, and other MCP-compatible assistants connect through a model-neutral interface instead of a captive embedded copilot.

SAFE

Make authority explicit

Roles, permissioned tools, approval gates, and action history keep AI useful without giving it invisible or unlimited control.

YOURS

Keep the operating layer portable

Provider choice, isolated deployments, versioned APIs, webhooks, exports, and an SBOM keep the platform inspectable and adaptable.

Conference operations, rethought for agents

See the operating system behind the event.

Open the live demo