Apple is taking a big step toward hands-off app creation.
With Xcode 26.3, Apple is officially introducing agentic coding, letting AI agents from OpenAI and Anthropic take on complex development tasks directly inside Xcode. Instead of just suggesting code, these agents can now act on their own to build, test, and refine apps.
From Suggestions to Autonomous Action
AI models like Claude and ChatGPT have been part of Xcode since Apple introduced intelligence features in Xcode 26. Until now, though, they were limited to assistance and couldn’t actually do things on a developer’s behalf.
That changes with Xcode 26.3.
By enabling an AI coding assistant, developers can let agents access deeper Xcode features to work toward a defined project goal. Apple says it worked closely with OpenAI and Anthropic to tailor their agents specifically for Xcode workflows.
These agents can now:
- Create new files
- Examine the structure of an existing project
- Build projects and run tests
- Review image snapshots to verify UI work
- Access full Apple developer documentation designed specifically for AI agents
Adding an agent is as simple as a single click in Xcode’s settings. Agents update automatically as new versions roll out, though developers will need an OpenAI or Anthropic account and will pay usage-based API fees.
Built for Flexibility and Efficiency
Apple says it focused heavily on performance, ensuring Claude Agent and Codex run efficiently with reduced token usage. Developers can also switch between agents within the same project, choosing whichever tool fits the task best.
And while Apple partnered directly with OpenAI and Anthropic, the feature isn’t locked to them. Xcode 26.3 supports any agent built on the open Model Context Protocol, and Apple is releasing documentation so developers can configure and connect their own MCP-compatible agents.
How Agentic Coding Works in Practice
Using natural language, developers can describe what they want done, and Xcode handles the rest. Apple outlines the workflow like this:
- A developer asks an integrated agent to add a new feature to an app.
- The agent reviews the project structure.
- It checks relevant documentation, code samples, and the latest APIs.
- The agent writes and adds code.
- It builds the project and verifies the result.
- If errors or warnings appear, the agent iterates using build logs until everything is resolved.
- The agent finishes by delivering a full summary of what it changed.
Developers can watch everything unfold in real time. Xcode shows a transcript of the agent’s actions in the sidebar, lets developers jump directly to newly added code, and even roll back changes if needed to undo results or try a different approach.
Productivity, Learning, and Creativity
Apple sees agentic coding as more than just a speed boost. The company says it can also help developers learn new APIs, discover better patterns, and experiment with ideas faster than before.
“Agentic coding will allow developers to simplify workflows, make changes quicker, and bring new ideas to life,” Apple says, positioning the feature as both a productivity tool and a learning aid.
Susan Prescott, Apple’s vice president of Worldwide Developer Relations, put it this way:
“At Apple, our goal is to make tools that put industry-leading technologies directly in developers’ hands so they can build the very best apps. Agentic coding supercharges productivity and creativity, streamlining the development workflow so developers can focus on innovation.”
When You Can Try It
The release candidate for Xcode 26.3 is available to developers starting today, with a full public release expected within the next week or so.
If this rollout goes as planned, Xcode is about to feel a lot less like a code editor and a lot more like a collaborative AI teammate.
