OpenAI just launched Dots. Let’s understand what they are and the architecture behind them. For most people, AI has been something you open, ask a question, and close. Dots are built to keep working after you’ve closed the tab: they take on a responsibility, run on their own cloud computer, and keep moving it forward while you’re offline.
To see how that works, it helps to think of a theatre.
Components of Dots Architecture
Think about what makes a theatre different from someone reciting lines in their living room. It has a building, a crew, a calendar of performances, and a prompt book that holds every cue. The show exists whether or not the author is watching.
In technical terms, a Dot is a persistent AI agent with its own runtime. OpenAI has given each Dot access to its own cloud computer (VM) and browser. They run on GPT-6 Astra and connect to over 4,000 apps. The agent’s context and tools stay put between conversations, so the work carries on after the chat ends. OpenAI described giving a Dot a responsibility, such as monitoring bug reports, preparing a budget cycle, or moving an application off a retiring API, and letting it work through the steps.
Before I breakout the architecture visually some thing to share. I’m running a live, 60-minute session on this exact topic: Build an Agent Harness Like Claude Code, October 20, 2026, 11:00 AM.
