A dot’s cloud computer
OpenAI says each dot has its own cloud computer and browser, a sandbox OpenAI maintains, and plugin permissions you choose.
This guide restates “Dots can work on their own” on Introducing dots and the workspace, sandbox, and Plugins material on How we build safety, security, and privacy into dots. It does not invent operating-system details beyond those pages.
Its own computer and browser
On Introducing dots, OpenAI says a dot has its own cloud computer and its own browser.
On Introducing dots, OpenAI says dots can connect to over 4,000 apps via plugins.
On Introducing dots, OpenAI says you can open the dot’s computer to inspect work.
Other devices stay separate unless you connect them
On Introducing dots, OpenAI says a dot can connect to other devices in addition to its own.
On Introducing dots, OpenAI says connecting your laptop is optional.
On Introducing dots, OpenAI says your computer and its contents stay separate unless you choose to connect it.
On How we build safety, security, and privacy into dots, OpenAI says connecting your computer is also subject to the local sandbox and action checks.
On that safety post, OpenAI says microphone or camera access needs the device permissions you grant to ChatGPT.
Several projects at once
On Introducing dots, OpenAI says a dot can run a project while working on several others.
On Introducing dots, OpenAI says you can bring new tasks without separate threads or directing every step.
A protected workspace
On How we build safety, security, and privacy into dots, OpenAI says sandboxing restricts what code and tools that dot can access.
On that safety post, OpenAI says it isolates users’ cloud environments from one another and maintains the underlying Linux operating system and Chrome browser.
On that safety post, OpenAI says dots run code in environments separate from the systems that coordinate work and enforce safeguards.
On that safety post, OpenAI says they cannot use that access to change those safety systems or turn off required checks.
Plugins and permissions
On How we build safety, security, and privacy into dots, OpenAI says you choose which apps to connect.
On that safety post, OpenAI says permissions are the existing ChatGPT app controls, shared across ChatGPT, ChatGPT Work, and Codex.
On that safety post, OpenAI says that in Plugins settings you can review connected accounts.
On that safety post, OpenAI says disconnecting an app stops new sharing through that connection, but information a dot has already learned remains in its own context.
On that safety post, OpenAI says permissions do not override mandatory safety requirements.
Resetting context
On How we build safety, security, and privacy into dots, OpenAI says each dot has its own context that you can reset.
What this page does not add
This page does not invent extra OS versions, menus, or setup minutes. For confirmations and Activity View, see Staying in control of a dot. For Auto-review, see Auto-review and approvals.