OpenAI’s Dots Are Always-On AI Agents—and Its Answer to Meta’s Muse
wired ·
OpenAI has officially launched Dots, a new suite of always-on AI agents introduced at its DevDay 2026 event in San Francisco. Powered by the GPT-6 Astra model, these personalized agents are designed to proactively crawl the web and execute multi-step tasks on behalf of users. Unlike traditional chatbots, Dots maintain continuous context by connecting to various apps and learning user preferences over time. The system allows users to manage tasks such as planning meals or launching websites through interfaces like ChatGPT, Slack, and Microsoft Teams. Access is currently restricted to ChatGPT Pro subscribers paying $100 per month, with plans to eventually support multiple agents simultaneously. To ensure safety, Dots require explicit approval for sensitive actions and offer customizable rules for user-defined boundaries. This release positions OpenAI directly against Meta’s recently popularized Muse agent, aiming to capture the growing market for personal AI assistants.
wired