Skip to main content
Back to News
news/AI Agents

OpenAI Codex Record and Replay: Teaching AI Your Workflows by Example

Record and Replay lets people demonstrate repetitive computer work once, then reuse that sequence as a skill that Codex can repeat later.

Stefan Trbojevic

Stefan Trbojevic

20 June 20263 min read
LinkedIn

The takeaway

The automation interface is shifting from written instructions to observed examples.

Why it matters for builders

The automation interface is shifting from written instructions to observed examples.

OpenAI has introduced Record and Replay inside Codex — a feature that lets you teach the AI repetitive tasks by simply doing them once on camera.

HOW IT WORKS

You hit record, perform a task on your computer — uploading a YouTube video, filling out a spreadsheet, extracting data from a tool — and type done. Codex watches the recording, understands the workflow, and creates a reusable skill. Next time, you drop the files into the chat and say: Upload this video using the YouTube upload skill.

Codex repeats every step exactly: selecting files from your computer, uploading them, setting privacy options, copying metadata — the full sequence preserved.

THE IMPLICATION

This is not just a convenience feature. It changes the abstraction layer between humans and automation. Instead of describing a process in words or writing scripts, you demonstrate it. The AI learns by observation, not instruction. For knowledge workers with repetitive digital tasks — content creators, marketers, operations teams — this collapses the gap between having a workflow and automating it.

KEY TAKEAWAY
Record and Replay shifts automation from write instructions to show example. The barrier to automating a digital task just dropped from knowing how to code or prompt-engineer to knowing how to do your job. That is a meaningful reduction in friction.

Share𝕏

The Automation Brief

Read 5 AI stories instead of 50.

The essential moves in AI agents, models, automation and infrastructure — filtered for builders and operators, with the part that actually matters.

No noise. Unsubscribe anytime.

Editorial notes

Reported by

Stefan Trbojevic

Edited by

n8n Lab Editorial

Published

20 June 2026

Updated

20 June 2026

Sources

Source links pending editorial review.

AI disclosure: AI assisted with research and drafting. Factual claims are reviewed by an editor.

n8n Lab is an independent service provider. We are not affiliated with, endorsed by, or sponsored by n8n GmbH. “n8n” is a trademark of n8n GmbH and is used here only to describe the platform-specific implementation and automation services we provide.