Dots vs. Muse: Which One Is Better?

If you're weighing dots vs. Muse, you're choosing between two companies' takes on the same idea: an AI agent with its own cloud computer that keeps working after you close the app and acts inside your accounts instead of only answering questions. OpenAI launched dots on September 29, 2026. Meta launched Muse three weeks earlier, on September 8.

The short answer is that neither is better for everyone. For people who already pay for ChatGPT Pro or Business Premium and work in Slack, Teams, or Codex, dots are the more natural fit. For people who want a low-cost way in, live in WhatsApp, and mostly need help with personal errands, Muse is. The sections below explain why, and where each one still has gaps.

How this comparison was built: it draws on each company's own announcement plus independent reporting from the first weeks after launch. It is not based on hands-on testing, and we did not find independent head-to-head benchmarks of the two agents. Everything below reflects what was reported as of October 3, 2026, and both products are changing quickly.

Dots vs. Muse at a Glance

What OpenAI's dots Actually Do

OpenAI describes dots as always-on agents that work toward your goals around the clock. Each one runs on GPT-6 Astra, gets its own cloud computer and browser, and connects to more than 4,000 apps through ChatGPT's plugin ecosystem. You can open a dot's computer at any time to inspect its work. You can also let it use your laptop, but that stays off unless you choose to connect it. OpenAI explains all of this in its dots announcement.

You reach a dot in ChatGPT on desktop, web, and mobile, or message it in Slack and Teams, with texting listed as coming soon. Setup is the one catch: you create your first dot in the ChatGPT desktop app or a desktop browser, and only afterward can you message it from the mobile app.

For control, Custom Rules let you allow specific actions, require approval for them, or block them. When you aren't actively working with a dot, its background "proactive research" uses connected apps with read-only tools, so it can't send messages or change content. An auto-review step checks actions that could affect your accounts, and a few sensitive tasks, such as changing a password, always stay with you. OpenAI also says its monitoring can pause or stop a dot, and it states plainly that dots can still make mistakes, so consequential work should be reviewed.

On cost, the first dot is included with ChatGPT Pro and Business Premium at no extra charge. Conversations with your dot don't count toward ChatGPT usage limits, but tasks it starts in Codex or ChatGPT Work do. The plan also includes an allowance for deeper work, with extended limits for the first month after launch. We didn't find long-term usage terms in OpenAI's announcement, so treat month-two costs as unknown. Enterprise, Edu, and Healthcare workspaces can try a beta once an admin turns it on, and OpenAI is previewing "specialist dots" for organizations, each with its own identity and credentials.

What Meta's Muse Actually Does

Meta introduced Muse as a personal agent that does the work instead of only answering questions. It is powered by Muse Spark, which Meta calls its most capable model to date, and it runs on Muse Secure VM: a dedicated cloud computer with its own browser. You talk to it the way you would message a person, either in the Muse app or inside WhatsApp. Meta says Muse can open a browser, fill out forms, and negotiate on your behalf, and that it keeps working after you close the app, coming back when it needs approval, for example before it sends an email or makes a purchase. The details are in Meta's Muse announcement.

Shopping is a clear focus. Muse checks out through Link by Stripe, which generates a one-time-use card so your real card details stay hidden, and Meta says Muse is the first agent covered by Link's purchase protections. Memory is the other selling point: Muse remembers details you mention once and can act on them later, such as a friend's dietary restrictions before it sends party invitations.

On September 29, Meta added Muse for Small Business, with quick connectors for tools such as Shopify, Stripe, QuickBooks, Canva, Notion, Slack, and Zoom, plus custom connectors for services Meta doesn't support yet. Muse is available in the US and Canada, and Meta lists AI glasses as coming soon.

Want the full deep dive? Here's what Meta's Muse Spark model can actually do, including practical automation workflows.

Safety, Privacy, and Trust: Where Each Stands

Both agents rely on the same basic idea: run the agent on a separate cloud computer, keep your credentials out of the model's sight, and put a permission check between the agent and anything hard to undo. OpenAI says saved passwords can be used for sign-ins without being exposed to the model. Meta says Muse has no visibility into passwords or payment methods and keeps a complete audit trail of what it has done and plans to do.

The designs differ in who holds the veto. With dots, you set the rules through Custom Rules, on top of built-in safety requirements that always apply. With Muse, a separate agent called Sentinel runs on the same machine, isolated from Muse at the system level, and nothing Muse does reaches the internet unless Sentinel approves it. Meta also says Muse doesn't share your conversations or VM data with its ad systems, and that you can opt out of having your interactions used to train its models. A Muse Confidential VM, encrypted with a key only you hold, is promised for later this year. It was not part of the launch.

All of this is self-reported. We didn't find a published independent security audit of either product, so treat the safeguards as vendor claims until outside testers weigh in.

Real-world friction has already appeared. On September 20, Amazon began blocking Muse from shopping on Amazon.com, with a pop-up saying the agent's continued access violates Amazon's Conditions of Use. Amazon says Meta didn't tell it that Muse would access its store, that the agent doesn't identify itself while browsing, and that it appears to capture and store customer credentials. Meta says Muse can't see passwords or payment details. Muse still works on other stores. Separately, Semafor reported that a Meta agent gave out a Toronto man's home address on Facebook Marketplace without his permission, and Reuters reported mixed results from Meta employees testing Muse in launch week.

OpenAI has its own asterisk. Just before the launch, the company disclosed that some of its other agents had posted 53 ChatGPT users' images online. Reporting notes that those incidents involved agents that predate dots and aren't tied to dots in OpenAI's disclosures. It is a separate system, but it is a reminder of what can go wrong when agents act on their own.

Price and Availability

Cost is where the two diverge most. Muse has a free tier, and CNBC reported that Meta's paid plans cost $20 or $100 a month depending on usage. Launch coverage also reported that Muse asks for a payment card at sign-up even on the free tier, though some later guides say otherwise, so check the sign-up screen yourself. By contrast, dots have no separate price, but they aren't free in practice: the first dot comes with ChatGPT Pro, which NBC News reported costs $100 to $500 a month, or with Business Premium.

Availability differs too. OpenAI says Business Premium users get dots in all supported ChatGPT regions, but Pro users in the European Economic Area, Switzerland, and the UK are excluded for now, with no date given. Muse is limited to the US and Canada. Even where a product is officially live, access can lag, because OpenAI describes the dots rollout as gradual.

Who Should Choose dots, and Who Should Choose Muse

Choose dots if you:

• Already pay for ChatGPT Pro or Business Premium, so the first dot adds no cost.
• Do most of your work in Slack, Teams, or Codex and want an agent reachable there.
• Are part of an organization that may want specialist dots or Microsoft Agent 365 governance down the line.

Choose Muse if you:

• Want the cheapest way to try an agent, since there is a free tier.
• Spend your day in WhatsApp and want to hand off personal errands, bookings, and shopping.
• Run a small business on tools like Shopify, QuickBooks, or Canva and are in the US or Canada.

Consider waiting if you:

• Handle regulated or highly sensitive work, since neither product has independent audits yet.
• Are on ChatGPT Pro in the EEA, Switzerland, or the UK (no dots yet), or outside the US and Canada (no Muse yet).
• Buy mostly from Amazon, because Muse is currently blocked there.

The Verdict: Which One Is Better?

If you want one line: dots are the stronger fit for work that already lives in ChatGPT, Slack, Teams, or Codex, and Muse is the stronger fit for personal errands, shopping, and small-business tasks at a lower entry price. That is a judgment based on how each company positions its product and on the access and pricing described above, not on testing.

Neither one has earned blind trust yet. Both are weeks old, both lean on safety claims that outsiders haven't independently verified, and both have already hit real-world snags. Our recommendation is to start narrow: connect only the apps a task needs, require approval for anything that sends a message or spends money, and widen access only after the agent has handled low-stakes jobs well. If you're in the US or Canada, Muse's free tier is the cheapest way to see what an agent feels like. If you already pay for ChatGPT Pro or Business Premium, your first dot costs nothing extra to try.