OpenAI adds scheduled agent execution to beta Agents API

OpenAI added scheduled execution to its beta Agents API on October 23, allowing developers to start agent tasks automatically, according to a report citing its community announcement and developer documentation. Developers can schedule recurring tasks with a time zone or specify a future time for a single execution. Each execution can create a new session or send a follow-up to an existing session that retains earlier context. An optional tool also lets agents manage schedules within their current session. Scheduled executions include a small random delay. Delivery means a session has received input or a notification; it does not mean the task has finished. Applications must handle presenting results and notifying users.