OpenAI's Codex now watches your screen to remember what you're working on

2026-04-20 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Cybersecurity & Data Privacy · Depth: Intermediate, quick

Summary

OpenAI has introduced Chronicle, a new feature for its Codex application, designed to enhance user experience by providing persistent context. Chronicle operates by recording the user's screen in the background, with AI agents converting these recordings into summarized Markdown files stored locally. These summaries serve as "memories" that Codex can access to understand ongoing projects, tools in use, and user references, eliminating the need for repeated explanations. The screen recordings are temporary, deleted after six hours, and the feature is currently an opt-in preview for ChatGPT Pro subscribers on macOS, excluding the EU, UK, and Switzerland. OpenAI advises users that Chronicle can consume API rate limits rapidly, poses risks of prompt injection attacks via displayed content, and stores memories unencrypted on the device.

Key takeaway

For AI architects and developers integrating OpenAI's Codex, you should carefully evaluate Chronicle's implications for data privacy and security. While it offers enhanced contextual understanding, the unencrypted local storage of memories and increased prompt injection risk necessitate robust security protocols. Consider its impact on your API rate limits and ensure compliance with regional data regulations, especially given its unavailability in the EU, UK, and Switzerland.

Key insights

OpenAI's Chronicle uses screen recordings to build local, contextual memories for its Codex app.

Principles

• Context persistence improves AI utility.
• Local storage can enhance privacy control.
• Opt-in features manage user adoption.

Method

AI agents process background screen recordings into summarized Markdown files, which are then stored locally as memories for the Codex application to reference.

In practice

• Enable "Memories" and "Chronicle" in Codex settings.
• Grant macOS screen recording permissions.
• Monitor API rate limit consumption.

Topics

• OpenAI Codex
• Chronicle Feature
• Screen Recording
• AI Contextual Memory
• Prompt Injection

Best for: CTO, AI Architect, VP of Engineering/Data, AI Engineer, Software Engineer, AI Security Engineer

Related on AIssential

• OpenAI's Codex can now watch you work once and repeat the task forever — OpenAI's Codex enables AI agents to learn and automate complex workflows through single-demonstration recording.
• When Should Memory Stay Silent: Measuring Memory-Use Boundaries in Memory-Augmented Conversational Agents — LLMs often inappropriately integrate sensitive user memory, requiring explicit boundary management at retrieval and generation.
• Why Local-First AI Agents Are the Future (And Why It Matters for Your Privacy) — AI agents, while increasingly capable of deep system access, raise critical privacy concerns as most cloud-based architectures upload sensitive data like full screen screenshots, passwords, and financ…
• New Codex features include the ability to use your computer in the background — OpenAI's updated Codex app introduces background task execution and an in-app browser, advancing its "super app" vision.
• Governed Memory: A Production Architecture for Multi-Agent Workflows — Governed Memory provides a shared memory and governance layer for multi-agent systems, enhancing recall and reducing token usage.
• We have built a system that prevents context contamination of AI by using periodic memory updates. — Periodic memory updates and context resets can prevent AI context contamination and improve reliability.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.