- OpenAI Dots mobile brings always-on AI agents to your phone, working 24/7 while you sleep
- Each Dot has its own cloud computer and browser, powered by GPT-6 Astra
- Dots are included in the Pro plan with no usage drawdown at launch
- Mobile users can delegate calls, calendar checks, and overnight triage to named Dots
- Start with one recurring job before handing over high-stakes tasks
What OpenAI Dots Mobile Actually Is
OpenAI Dots mobile represents a shift from chatbots you prompt to persistent agents you manage. Announced at the September 29, 2026 DevDay keynote, Dots are always-on agents that keep working after you close the app or lock your phone. Each Dot gets its own cloud computer with its own browser, runs on the GPT-6 Astra flagship model, and remembers your preferences between sessions.
On mobile, the experience centers on delegation rather than typing. OpenAI's own launch promo shows two users handing ongoing responsibilities to their Dots on phone and desktop — tasks like introducing a new hire or managing a registration call. You are not retyping prompts every session; you are assigning a job description once and letting the Dot show up daily.
Video Highlights:
- Dots launched at DevDay 2026, powered by GPT-6 Astra
- Named agents (Alfred, Felipe, Jojo) with persistent memory between sessions
- Real-world demo shows a Dot making a live phone call to a tennis club
- Included in the Pro plan with no usage metering at launch
- Approval gates required for actions with real consequences
Stop thinking of Dots as software you use. Think of them as staff you manage. A contractor does one favor; an employee with a job description compounds value daily. That framing changes how you assign mobile tasks.
Core Mobile Features Explained
The mobile Dots experience is built around five capabilities demonstrated on launch day. Together they turn your phone into a management console for agents running in OpenAI's cloud.
| Feature | What It Does | Mobile Use Case | Risk Level |
|---|---|---|---|
| Phone calls | Holds real voice conversations with humans | Registrations, bookings, errands | High |
| Contract review | Reads paperwork, flags errors before they cost money | Approving documents on the go | High |
| Calendar monitoring | Watches your schedule, suggests changes | Dinner planning, meeting triage | Low |
| Overnight triage | Reads updates, summarizes what matters | Morning briefing at breakfast | Low |
| Remote computer access | Edits video, draws on a remote machine | Creative tasks while away from desk | Medium |
The standout mobile demo comes from early-access creator Roberto Nixon: while he cooked dinner, his Dot was on the phone with a tennis club, completing a real registration call with a real human. In the same walkthrough, the Dot reviewed contract paperwork and caught a mixup before it became a costly problem.
Persistence
- Dot remembers between sessions
- Learns your work preferences
- Shows up daily without re-prompting
Own Cloud Computer
- Dedicated browser per Dot
- Works while your phone is idle
- No local resources required
Named Agents
- Alfred, Felipe, Jojo style identities
- Ongoing responsibilities, not one-off tasks
- Manageable like team members
The always-on promise lives or dies on reliability. A Dot that works overnight is only useful if you trust what it did while you were not watching — and that trust takes weeks of clean runs, not one keynote demo.
Mobile Setup: Step-by-Step
Getting started with Dots on mobile follows a deliberate path. Resist the urge to hand over five jobs on day one.
Confirm Plan Access
Dots are included in the ChatGPT Pro plan per launch materials, with no usage drawdown announced. Verify the Dots feature appears in your mobile app after rollout reaches your account — day-one availability was limited.
Define One Recurring Role
Write a single bounded job description, such as "every morning, read overnight AI news and flag the three stories that matter." Recurring and bounded tasks are the perfect first Dot job.
Keep Approval Gates On
Leave sign-off rules enabled for anything with real consequences — calls, contracts, payments. Launch coverage consistently flags approval rules as the default safety posture.
Review the Morning Summary
Check the overnight briefing your Dot prepares. If the summary is boring and accurate, that is success. Boring is the goal.
Gradually Expand Scope
Only delegate contract review or phone calls after the Dot has proven itself on low-stakes recurring work and caught its first mistake correctly.
The best first job is recurring, bounded, and low-cost if it fails. A bad news summary costs you nothing; a bad contract decision costs real money. Dots earn the scary jobs the way every employee does — by nailing the boring ones first.
Dots vs Meta Muse: Mobile Comparison
The comparison everyone is making is OpenAI Dots versus Meta's Muse. Both are always-on agents, but they bet on opposite visions of what an AI assistant becomes.
| Dimension | OpenAI Dots | Meta Muse |
|---|---|---|
| Core bet | Employee with a job description | Companion across your apps |
| Where it lives | OpenAI cloud, own computer per Dot | Meta's apps and your devices |
| Mobile model | Delegate work, review results | Companion woven into daily app use |
| Pricing at launch | Bundled into Pro plan, no metering | Tied to Meta ecosystem |
| Working style | Works overnight while you sleep | Present where you already hang out |
| Boldest feature | Real phone calls + contract error catching | Deep app integration |
An employee does work you would rather not do. A companion keeps you company while you do it. OpenAI is making the bolder product bet, because persistence plus a flat Pro-plan bundle is how you change daily behavior rather than just demo well.
Side-by-side testing of both products has not been completed yet. Treat this as launch-day framing based on official materials, not a tested verdict.
Mobile Readiness Checklist
Before trusting a Dot with meaningful mobile work, run through this readiness list.
Before You Delegate:
- Confirm Dots access on your Pro account
- Write one bounded recurring job description
- Enable approval gates for calls and contracts
- Review morning summaries for one full week
- Verify the Dot correctly flags at least one issue
Known caveats worth tracking as rollout continues:
| Concern | Why It Matters | Current Status |
|---|---|---|
| Day-one coverage only | Edge cases and rough edges unfilmed | Early access limited |
| Reliability unproven | Overnight trust requires clean runs | Weeks of data needed |
| Safety line placement | Asking permission for everything feels like a chatbot | Approval rules expected default |
| Platform integrations | HiField plugin puts creative crews in ChatGPT | First version likely rough |
| Legal questions | Agents with browsers and phones worry lawyers | On-site DevDay discussion ongoing |
Do not hand your Dot contract authority until it has demonstrated it can catch a mistake. The value of an agent is not the errands it runs — it is the mistakes it stops. That judgment takes evidence.
FAQ
Q: What is OpenAI Dots mobile?
OpenAI Dots are always-on AI agents that work 24/7 from OpenAI's cloud, each with its own cloud computer and browser, powered by GPT-6 Astra. On mobile, you manage named Dots that handle ongoing responsibilities like calls, calendar monitoring, and overnight triage rather than one-off prompts.
Q: Do I need a separate subscription for Dots?
No. According to launch materials from September 2026, Dots are included in the ChatGPT Pro plan with no usage drawdown — no credit metering or top-ups were announced at launch.
Q: Can a Dot really make phone calls?
Yes. Early-access creator demonstrations showed a Dot holding a real phone conversation with a tennis club to complete a registration while its owner cooked dinner. Actions with real consequences are expected to require your sign-off first.
Q: What should I delegate to a Dot first?
Start with one boring, recurring, bounded job — such as reading overnight news and flagging the three most important stories. If it fails, the cost is a bad summary, not a bad contract. Expand scope only after the Dot proves reliable.