OpenAI launched Dots, a new line of "always-on" AI agents, at its DevDay conference on Tuesday, positioning the product as a direct competitor to Meta's Muse (released September 8) and SpaceXAI’s Grok Bot (released August 11).
Powered by the GPT-6 Astra model, each Dot runs on a dedicated cloud computer with a web browser and connects to over 4,000 apps, allowing it to autonomously execute tasks like booking flights or managing code in the background without constant user prompts.
Dot can run scripts, research topics, download files, and complete web-based tasks on its own. Optional access to your local machine via the ChatGPT desktop app is available if configured.
Unlike traditional chatbots, Dots can be customized with unique names and avatars, and are designed to work continuously toward user-defined goals, learning preferences over time and integrating with platforms like Slack, Microsoft Teams and mobile and desktop voice calls, maintaining continuous context across platforms.
Dots differ from earlier OpenAI tools like Operator by featuring a persistent, always-on environment and broader app connectivity.
The rollout initially targets ChatGPT Pro, Business Premium, and Enterprise subscribers, with the first Dot included at no extra cost.
The launch follows recent safety scrutiny for OpenAI, including the shelving of the GPT-6.1 Astra model; consequently, Dots underwent additional safety testing and feature read-only background modes, and mandatory user approval for sensitive actions like changing passwords or deleting data, to prevent unauthorized actions, according to the company.
SpaceXAI trolled OpenAI on X following Dots announcemnet Tuesday. The Elon Musk's SpaceX AI division apparently bought the dot.com domain which redirects to the Grok Bot website.