AI_SLANG_ENTRY

What Are Dots?

Dots are OpenAI's always-on AI agents that can learn from feedback, work toward ongoing goals, and use a cloud computer plus connected apps to complete work over time.

AI_TASTE=2/5 ██░░░ TREND=HIGH

What does Dots mean?

Dots are OpenAI's always-on AI agents that can learn from feedback, work toward ongoing goals, and use a cloud computer plus connected apps to complete work over time.

A dot is an OpenAI agent intended to keep working between conversations. It has a cloud computer and can use approved connected apps, while the person can check in, give feedback, and steer the result.

Origin and usage

OpenAI introduced dots in September 2026 as a product category for persistent agents that work on a person's behalf beyond one chat session.

Source type: product-term. Last checked: 2026-10-06.

Product term verified against OpenAI's launch announcement. Dot capabilities, availability, connected apps, work limits, pricing, and model access are product details that may change.

Primary reference

Dots versus ordinary AI chat

An ordinary AI chat starts with a request and returns an answer. OpenAI describes dots as agents that can keep context, work toward an ongoing goal, and report back while the person focuses elsewhere.

The distinction is about the surrounding agent system: persistence, tools, feedback, and a cloud computer—not simply about a model being able to generate text. Dots are not a synonym for every AI assistant.

Why a cloud computer matters

OpenAI says a dot has its own cloud computer. That gives the agent an environment in which to use approved tools and connected apps instead of only drafting instructions inside a chat window.

A cloud computer does not remove the need for access controls. The work a dot can perform depends on the connections, permissions, and review steps available in the current product.

What can change

  • Product availability, connected apps, and limits can change.
  • Connected apps, plugins, and allowed actions.
  • Work-speed and monthly-capacity limits.
  • Models, pricing, and approval controls.

Examples

  • My dot gathered the account history before the customer call and left the open questions for me.
  • Instead of a one-off prompt, I gave the dot a goal and reviewed its work as it progressed.

FAQ

What are Dots?

Dots are OpenAI's always-on AI agents. OpenAI says they can learn from feedback, use a cloud computer, and work toward goals over time.

Are dots just ChatGPT conversations?

No. OpenAI presents dots as persistent agents that can work through connected apps and a cloud computer, rather than only responding to a single chat prompt.

Can a dot access every app automatically?

No. Availability, connected apps, permissions, and approval controls depend on the current product and the access a person or organization chooses to grant.

Further reading

Related AI slang