OpenAI introduces Record & Replay plugin for Codex to automate workflows
The new macOS feature lets users demonstrate tasks instead of writing prompts, turning everyday workflows into reusable automation skills
OpenAI just made automation a lot more visual. The company’s new Record & Replay plugin for Codex lets users simply perform a task on their Mac, and the AI watches, learns, and packages it into a repeatable skill.
A user initiates the Record & Replay plugin, performs their workflow as they normally would, and Codex captures the entire sequence. The AI then converts that recorded demonstration into what OpenAI calls a “skill,” a reusable automation that can be triggered again without manual intervention. Filing expenses, booking parking spots, publishing videos: instead of describing these workflows in textual prompts, you just do them once while Codex takes notes.
The feature is currently exclusive to macOS users.
Once a skill is created, teams can share these developed skills broadly, meaning one employee’s recorded workflow can become the entire department’s automation.
Codex’s growing plugin ecosystem
At its initial launch, Codex supported more than 20 third-party plugins with integrations spanning major applications like Slack, Gmail, Notion, Salesforce, and HubSpot.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
In March 2026, OpenAI expanded the plugin ecosystem further to support multiple platforms, including CLI and VS Code.
By early June 2026, OpenAI introduced role-specific plugin bundles tailored for distinct functions like sales and analytics.
What this means for investors
This announcement makes no references to cryptocurrencies, tokens, or blockchain technology anywhere in the Codex developments.
For investors focused purely on tech and productivity innovation, Codex’s trajectory from plugin marketplace to role-specific bundles to visual workflow capture represents a coherent product strategy aimed at capturing enterprise budgets that currently go toward manual processes and legacy automation tools.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
More in AI
See all-
Exin Therapeutics lets an AI agent run a mouse brain injection
-
Anthropic reveals nearly $900 billion in future commitments in IPO filing
-
Anthropic deploys Claude agents to stress-test training data monitors
-
Anthropic’s AI now handles 26% of core R&D work, report says
-
OpenAI trims its funding goal by $10 billion as AI investor demand cools
-
Cohere opens Series E targeting up to C$7 billion, leans on neocloud providers for compute

