dots glossary
The terms that show up in any explanation of dots, defined in one sentence so you can come back whenever you need them.
- dot #
- An always-on OpenAI agent that works on your behalf continuously. It has its own cloud computer, its own browser, and credentials for the systems you authorize.
- GPT-6 Astra #
- OpenAI’s frontier model, which dots run on. The model reasons and decides; the dot is the agent that executes.
- dot's computer #
- The isolated cloud environment where a dot does its work: files, programs and code. You can open it at any time to inspect what the dot did.
- proactive research #
- The mode a dot enters when you are not actively working with it. It looks for ways to help using your connected apps, with tools limited to read-only operations.
- auto-review #
- A check that runs before an action that could affect your accounts or share information. It verifies the action complies with your instructions, custom rules and safety requirements.
- custom rules #
- Rules you write to allow, require approval for, or block specific dot actions. They sit on top of built-in rules, which are never disabled.
- plugin ecosystem #
- The set of integrations that lets dots connect to more than 4,000 applications.
- specialized dot #
- A dot with its own identity, credentials and responsibilities inside an organization. It is configured for a specific area such as procurement or customer service.
- usage quota #
- The budget for deeper work included in your plan, separate from ChatGPT usage limits. Conversations with your dot do not consume ChatGPT limits.
- enabled markets #
- The list of countries where OpenAI makes a product available. Availability depends on the market, not the interface language.
- activity view #
- The panel where you follow a dot’s progress, including background work, and give it new direction.
- always-on agent #
- A product category that does not wait for a prompt: it runs in the background, has its own computer, and pursues goals continuously. dots, Muse and Grok Bot are the three main ones of 2026.
- Muse #
- Meta’s always-on agent, introduced in early September 2026. It is dots’ direct competitor, not an alternative name for OpenAI’s product.
- Grok Bot #
- xAI’s always-on agent, launched in August 2026. It uses one cloud computer per user, shared across all their bots, unlike dots, which gives each agent its own computer.
- GPT-6.1 Astra #
- A later version of the model that powers dots, which OpenAI decided not to ship over safety concerns reported on 28 September 2026. Not to be confused with GPT-6 Astra.
- Astra Ultrafast #
- The higher-speed mode included with ChatGPT’s Pro 500 tier, the $500 per month plan.
- prompt injection #
- A technique where a website, email or document carries instructions to trick the agent into doing something you did not ask for. Encountered content does not grant permission on its own.
- EEA #
- European Economic Area: the 27 European Union member states plus Iceland, Liechtenstein and Norway. One of the markets excluded from the dots Pro plan.
- Microsoft Agent 365 #
- A Microsoft platform that OpenAI is integrating specialized dots with, so enterprises can manage them using governance and security controls they already have.