OpenAI has rolled out a major update to Codex, transforming it from a simple coding assistant into a comprehensive workplace tool. The new version is designed to streamline workflows, automate repetitive tasks, and enhance developer productivity across the entire software development lifecycle.
With this upgrade, Codex can now interact more directly with a user’s computer environment. It is no longer limited to generating code; it can operate desktop applications, access various tools, and perform tasks in the background without interrupting ongoing work. This marks a shift toward AI systems that actively assist in daily operations rather than just responding to user prompts.
One of the most notable features is its new memory capability. Codex can retain context from past interactions, including user preferences, corrections, and frequently used inputs. This enables it to deliver more accurate and personalized results over time. By learning from previous actions, Codex minimizes repetitive inputs and speeds up task completion, particularly in complex development scenarios.
Beyond memory, Codex now supports proactive task management. It can recommend next steps based on current projects and even schedule tasks independently. It can resume work on long-term assignments automatically, continuing tasks over several days or weeks. This turns Codex into a semi-autonomous assistant that helps maintain continuous progress with minimal supervision.
The update also allows multiple AI agents to run simultaneously on a single device, especially on macOS. These agents can collaborate on tasks such as frontend development, application testing, and workflows involving tools without APIs. They operate independently to ensure they do not interfere with other applications, maintaining a smooth user experience.
To improve usability, Codex now includes a built-in browser. This allows users to interact directly with web pages by adding comments or instructions that the AI can understand and execute. This feature is particularly useful for frontend development, where visual feedback and contextual changes are essential.
Another enhancement is integrated image generation using an advanced model. Developers can create visuals, mockups, and design elements within the same workspace where they write and test code. By combining screenshots, generated images, and code, teams can quickly prototype ideas and refine product concepts without switching between tools.
Codex also offers deeper integration with development workflows. It now supports responding to GitHub review comments, managing multiple terminal sessions, and previewing files such as PDFs, spreadsheets, and presentations directly within the interface. A new summary panel provides a clear overview of agent activities, including plans, sources, and outputs, improving visibility into ongoing tasks.
These updates are being rolled out to users of the Codex desktop application who are signed in through ChatGPT. Personalization features, including memory and context-aware suggestions, are expected to expand to enterprise and educational users, with broader regional availability planned. Advanced computer control features are currently available on macOS, with further expansion expected.
Overall, this update highlights the evolution of Codex into an intelligent and autonomous system that actively collaborates with users. By combining memory, automation, and multi-agent functionality, Codex is becoming a powerful productivity tool for modern development and digital workflows.
Read more such articles from our Newsletter here.
