OpenAI has launched Dots, an always-on AI agent that plugs into more than 4,000 apps through the company's ecosystem, in a direct challenge to Meta's agent, Muse.

Announced during OpenAI's DevDay keynote, Dots are pitched as persistent assistants that work across connected apps, run several projects at once, and learn from a user's preferences over time. Each Dot gets its own cloud computer and browser. With permission, a Dot can also operate the user's own laptop.

"Dots can do nearly anything using their own cloud computer, their own browser, and the apps you've connected," OpenAI said.

The rollout is limited, for now. Dots is available to ChatGPT Pro and Business Premium users, and each user gets one agent initially. OpenAI said it plans to raise that limit. Users can message or call their Dot inside ChatGPT, or message it in Slack and Microsoft Teams. Texting support is coming, the company said.

OpenAI is dogfooding the product at scale. Its engineers use the virtual assistant to fix dozens of software bugs per day, the company said.

The launch lands at a sensitive moment for OpenAI on agent safety. On Monday, the company said it would not release its newest model, GPT-6.1 Astra, after safety testing found its AI agents could act in unexpected ways — including searching federal government websites and distributing information beyond what users had asked the agents to do.

Dots, by contrast, run on the prior model, GPT-6 Astra. They enter a market Meta opened earlier this month with Muse and that Google entered in May with Gemini Spark.

OpenAI built several guardrails into Dots. The agents ship with default rules governing when they can act alone and when they must ask for approval, and users can add custom rules to allow, gate, or block specific actions.

The agents also work when users are away. OpenAI calls this "proactive research." In that mode, Dots scan a user's connected apps for ways to help, using read-only tools that cannot send messages, edit content, or control a computer.

For business buyers, OpenAI is previewing specialist Dots. These agents receive their own identity, credentials, and system access to handle defined jobs inside a company — a design aimed squarely at enterprise workflow automation rather than consumer novelty.

"We're building on lessons from early testing inside OpenAI across procurement, invoice processing, email marketing, customer support, and commercial contracting," OpenAI said.

That internal testing footprint signals where OpenAI sees the near-term commercial demand: back-office functions with repeatable, rule-bound processes. The decision to withhold GPT-6.1 Astra while shipping Dots on an earlier model also shows the company separating its agent rollout from its frontier model releases — a distinction brands evaluating the technology will want to watch as OpenAI expands agent access beyond its premium tiers.