OpenAI has released a feature that lets Codex run tasks automatically on a set schedule, turning the language model into a proactive assistant. By configuring schedules and triggers, developers can have Codex generate weekly reports, morning briefs, or file summaries without manual prompts.

The update follows a growing demand for low‑effort automation in AI workflows. Codex can now return to an ongoing conversation, carrying context from previous runs, which streamlines recurring projects like status updates or data clean‑ups. When run locally, the assistant works best while the host machine remains awake.

Practical applications span from generating a daily briefing from yesterday’s work to summarizing new files in a folder. Developers first chat with Codex to refine the desired output, then lock the interaction into an automation. The result is a repeatable, easy‑to‑review process that frees users from repetitive manual checks.