BreakthroughsThursday, April 16, 2026· 2 min read

OpenAI Supercharges Codex with Desktop Control, Memory, and Image Tools

Source: The Verge AI

TL;DR

OpenAI's Codex update gives the coding agent the ability to control desktop apps, work in the background with parallel agents, generate images, and remember past interactions. These features boost developer productivity and bring more practical, agentic automation to everyday workflows while accelerating competition that benefits users.

Key Takeaways

  • 1Codex can now operate desktop apps and run in the background without disrupting your work.
  • 2Multiple agents can work in parallel and Codex gains memory to recall past interactions.
  • 3New image-generation capabilities expand Codex beyond pure code into design and content tasks.
  • 4The update sharpens competition with Anthropic and is likely to accelerate practical agent development.

OpenAI expands Codex into a more capable, agentic teammate

OpenAI's latest Codex update turns its developer-focused assistant into a more practical automation platform: it can control desktop applications, run unobtrusively in the background, spawn parallel agents, generate images, and retain memories of prior interactions. These enhancements move Codex from a code-writing helper toward a multi-modal, persistent assistant that can help with real-world developer workflows.

The ability to operate apps on your computer means Codex can perform tasks that previously required manual intervention — from orchestrating builds and tests to manipulating local files or even interacting with GUI tools. Because agents can run in parallel and operate in the background, developers can delegate routine or long-running tasks while continuing other work.

Memory and image generation broaden Codex's usefulness. Memory lets Codex recall project context and user preferences across sessions, reducing repetitive setup and making interactions smoother. Built-in image-generation capabilities expand the agent’s role into design and documentation, enabling richer outputs without stitching together multiple tools.

  • Improved productivity: hands-off automation for local tasks and parallel workflows.
  • More practical agents: persistence and multi-modal outputs make Codex more helpful across the project lifecycle.
  • Healthy competition: updates driven by rivalry with Anthropic accelerate innovation that benefits developers and end users.

Get AI Wins in Your Inbox

The best positive AI stories delivered to your inbox. No spam, unsubscribe anytime.