The Release Candidate of Xcode 26.3 is rolling out as we publish this with brand new agentic coding capabilities. With the update, developers can connect to either OpenAI’s Codex or Anthropic’s Claude Agent to assist their development work.
The agent sits in Xcode’s sidebar where developers can use it to plan new features, implement them, and review the results. As developers work, the agent generates a transcript of its actions, which lets developers follow along and interact with it. For example, code snippets will appear in the sidebar that can be clicked to take developers directly to the spot in the file where the agent made a change. Code updates can also be simultaneously previewed. Plus, by building OpenAI and Anthropic’s agents into Xcode, the agents have the benefit of the latest Apple APIs and documentation, which should result in better, more modern code.
Xcode 26.3 also integrates with Model Context Protocol, the open standard used to connect coding agents to other tools. That will allow any MCP-compatible agents to work with Xcode, running inside or outside Xcode, which broadens the possibilities for developers even further. Having watched a short demo of Codex assisting in the development of features in Xcode, version 26.3 looks as though it has a lot of potential to make experimentation easier and to iterate more quickly on ideas.
I’ve been building a couple of different apps recently, mostly from Terminal, which works but isn’t ideal. There are lots of apps out there that try to solve the agentic coding problem, including Visual Studio Code and Codex, which I covered yesterday. However, having seen a demo of Xcode in action, it’s jumped to the top of my list to try. After all, it’s an app with which I’m already familiar, and with Apple’s documentation and API knowledge baked in, my hope is that it speeds up my workflow even further.
For developers who want to learn more, Apple is holding a code along session Thursday covering the update.
