Developer — turn feedback into tested fixes. You’re shipping an app. Your ChatGPT dots agent watches customer feedback for recurring requests, scopes smaller improvements and bugfixes, builds and tests them, and brings complete PRs with short demo videos. You keep building the next major feature while reviewing what the dot prepared on its cloud computer. Pair with for developers when Codex or ChatGPT Work handoffs matter — those tasks still count toward usage even if chat with the dot does not.
Community-hosted · attributed
Community-hosted demo (attributed): ChatGPT dots on a remote Linux — voice-call directed PR review. Not an official product SLA.
Community demo — PR review on a remote Linux (attributed). @huacnlee shows a ChatGPT dots session on a remote Linux: install Rust, clone a PR, start review — directed by a phone voice call. Treat as a community walkthrough of the cloud-computer / always-on developer use case, not a product SLA. Official: Introducing dots, for developers.
Product launch — revise as scope changes. You’re leading a launch. Your dot learns audience, positioning, and creative standards. When product scope shifts, it works through what that means for the story, revises launch materials, and prepares drafts of asset and doc changes for your review. You return to a plan shaped by the direction you set — not a finished ship without approval.
What stays human. Approvals on consequential changes, brand voice final say, and merge decisions. OpenAI’s control model (Custom Rules, Auto-review, action review) is the spine — see workflow deep-dive and safety.
Do not invent SKUs. Plan gates and quotas: eligibility, pricing. Official narrative source: Introducing dots. More sketches: research / sales / creator, use-case sketches index (prompts / 24/7 demos).