Developer

OpenAI Updates Codex to Control macOS Apps

OpenAI has updated its Codex coding system to control desktop apps on macOS, generate images, and retain memory from past tasks. The changes position it against Anthropic's Claude Code amid rising competition. Features roll out first to macOS users signed into ChatGPT, with expansions planned.

Neura News

Neura News

Neura Market Editorial

April 16, 20263 min read
OpenAI Updates Codex to Control macOS Apps

OpenAI Updates Codex to Control macOS Apps

OpenAI has enhanced its Codex agentic coding and development tool with several new capabilities. The updates allow Codex to access and operate desktop applications on users' computers. This move comes during heightened competition with Anthropic, especially after strong performance from Claude Code. OpenAI appears to be directing more efforts toward matching that progress.

Desktop App Control on macOS

Codex can now run apps on your Mac without disrupting other activities. It operates in the background and supports multiple agents running at the same time. Developers benefit from this for tasks like testing frontend modifications, running applications, or using software without public APIs, according to OpenAI's announcement.

A demonstration shows Codex playing games such as Tic Tac Toe through desktop control. The rollout begins today for users of the Codex desktop app who log in via ChatGPT. Availability starts with macOS, and OpenAI has not shared dates for Windows or Linux support. Users in the European Union will gain access shortly.

OpenAI, founded in 2015 by Sam Altman and others, leads in AI models like GPT series and ChatGPT. Its agentic systems aim to handle complex tasks autonomously. Codex builds on this by focusing on coding and development workflows.

Image Generation and New Integrations

Codex now supports creating and refining images using the gpt-image-1.5 model. It includes fresh plug-ins for services like GitLab, Atlassian Rovo, and the Microsoft Suite. Users get native web browsing inside the app, complete with tools to annotate pages and give exact directives to the agent.

The #1 Newsletter in AI

Stay ahead of the AI curve

The most important updates, news, and content — delivered weekly.

No spam. Unsubscribe anytime.

Task automation improves too. People can repurpose prior conversation threads. Codex can set up its own schedules, activate itself, and resume extended projects automatically.

Anthropic, started in 2021 by former OpenAI executives including Dario Amodei, released Claude models known for safety and coding prowess. Claude Code has gained attention for developer tools, prompting rivals like OpenAI to respond.

Memory Feature for Better Personalization

A new memory system lets Codex store details from previous interactions. It recalls user preferences, fixes, and hard-to-find data. This opt-in preview aims to speed up work and match results that once needed lengthy prompts.

OpenAI plans to release personalization to Enterprise, Education, and EU customers in the near future. Robert Hart, a London-based AI reporter who previously covered health, science, and tech at Forbes, reported on these developments.

These updates strengthen Codex as a comprehensive tool for coders. They address gaps in app interaction, creativity, browsing, and continuity. The focus on macOS first aligns with Apple's large developer base using Macs for software creation.

Related on Neura Market

More from Neura News

Developer

LangChain and NVIDIA Launch NemoClaw Deep Agents Blueprint

LangChain and NVIDIA have released the NemoClaw for LangChain Deep Agents blueprint, designed to help enterprises build open, governed agent systems. The blueprint combines LangChain Deep Agents Code, NVIDIA Nemotron 3 Ultra, and NVIDIA OpenShell runtime, enabling teams to tune agents for their workloads, run them securely, and optimize for quality, cost, and speed. In evaluations, Nemotron 3 Ultra with a tuned LangChain Deep Agents harness achieved an aggregate score of 0.86 at a cost of $4.48, roughly 10 times lower inference cost than the next closest performing model.

Jul 25·7 min read
Funding

Prentis AI Lab Co-Founded by Reid Hoffman, Marc Pincus Seeks $100M

Prentis, a new AI research lab co-founded by Ritankar Das, Reid Hoffman, and Marc Pincus, is in talks to raise $100 million at a $1 billion valuation. The startup focuses on computer use models that automate office workflows. It has already signed contracts worth up to $50 million with several customers and claims its Hive-32B model outperforms rivals like OpenAI's GPT-5.4 and Anthropic's Claude Opus 4.6 on key benchmarks.

Jul 24·4 min read