OpenAI's Unification Play: One App, Three Products, No Fanfare

In March 2026, OpenAI announced it would combine ChatGPT, the ChatGPT application, and OpenAI Codex into a single desktop application. The company did not give it a splashy new name. There was no keynote unveiling a “SuperApp.” The announcement came and went without the usual fanfare.

The April 16 Codex update showed what that unification actually looks like in practice: not a rebrand, but a quiet blurring of lines between chatbot, coding assistant, and web browser.

Codex becomes the everything agent

The April 16 update is less a feature release than a territory expansion. Codex can now operate in the background on Mac, opening any application on the desktop and carrying out operations with a cursor that clicks and types. Multiple agents work in parallel without interfering with whatever else the user is doing.

More revealing is the new in-app browser. Codex users can now issue commands to the agent for web applications, with OpenAI signaling plans to eventually let Codex “fully command the browser beyond web applications on localhost.” This is Atlas’s territory. The browser OpenAI launched six months ago is being absorbed into the coding tool from the inside out.

Add to this: a memory feature that recalls previous work sessions, image generation for mockups and visuals, and 111 plugin integrations including CodeRabbit, GitLab Issues, Slack, and Google Calendar. Codex is no longer a coding assistant. It is becoming the unified surface that the March announcement promised.

The Anthropic shadow

“There is currently a low-grade war between OpenAI and Anthropic over who can release the most convenient and powerful AI-coding tools and, so far, Anthropic seems to be winning.”

That framing comes from TechCrunch, and it explains the urgency behind OpenAI’s unification push. Claude Code has become, in the publication’s words, “the tool of choice for many businesses.” Some of Codex’s new powers—remote Mac control, desktop operation—mirror capabilities Anthropic shipped for Claude Code last month.

OpenAI’s response is not to out-feature Anthropic on any single axis. It is to offer integration that Anthropic cannot yet match: a browser (Atlas), a general assistant (ChatGPT), and a coding agent (Codex) that share context, share memory, and increasingly share the same window.

The company has shifted resources to match. TechCrunch notes that OpenAI has “more fiercely competed with Anthropic in recent months, with a focus on enterprise capabilities and a retreat from consumer tools like its social video app Sora 2.” The superapp is not a consumer play. It is an enterprise consolidation.

The day before: Agents SDK

One day before the Codex update, OpenAI released an update to its Agents SDK with new sandboxing capabilities. The SDK now lets enterprises build agents that operate in controlled environments, working with files and approved tools within defined workspaces.

Karan Sharma from OpenAI’s product team described it as making the SDK “compatible with all of these sandbox providers.” The feature is designed for “long-horizon” work—complex, multi-step tasks that require agents to persist across sessions.

This is the infrastructure layer beneath the unification. If Codex is the consumer-facing everything-app, the Agents SDK is what lets enterprises build their own version of it. Both point to the same thesis: AI workflows are moving from single prompts to persistent, multi-tool agents that need somewhere to live.

What to watch

The March announcement said “one desktop application.” The April updates show the pieces moving together. What remains unclear is when the combined app actually ships—and what happens to users who prefer Atlas as a standalone browser or Codex as a dedicated coding environment.

OpenAI’s current product page still lists ChatGPT, Codex, and Atlas as separate offerings. The unification is happening in capabilities before it happens in packaging. By the time the single app arrives, the merger may already be complete.