Cursor Launches a New AI Agent Experience to Take On Claude Code and Codex

Cursor announced Thursday the launch of Cursor 3, a new product interface that allows users to spin up AI coding agents to perform tasks on their behalf. The product, developed under the codename Glass, is Cursor’s answer to agent coding tools such as Anthropic’s Claude Code and OpenAI’s Codex, which have taken off with millions of developers in recent months.

“Over the past few months, our business has completely changed,” Jonas Nelle, one of Cursor’s engineering leads, said in an interview with WIRED. “A lot of the product that brought Cursor here is not as important going forward.”

Cursor increasingly finds itself competing with leading AI labs for developers and enterprise clients. The company pioneered one of the first and most popular ways for developers to code with AI models from OpenAI, Anthropic, and Google, making Cursor one of those companies’ largest AI clients. But over the past 18 months, OpenAI and Anthropic have launched their own agent coding products and started offering them through highly subsidized subscriptions, putting pressure on Cursor’s business.

While Cursor’s main product allows developers to code in an integrated development environment (IDE) and tap an AI model for help, newer products like Claude Code and Codex allow developers to hand off entire tasks to an AI agent, sometimes running multiple agents at the same time. Cursor 3 is the startup version of an agent-first coding product. According to Nelle, the product is optimized for a world in which developers spend their days “talking to different agents, monitoring them and seeing the work they’ve done,” rather than writing code themselves.

Cursor is launching its new agent coding interface within its existing desktop application, where it will live alongside the IDE. In the center of a new window in Cursor, there is a text box where users can type, in natural language, a task they want an AI agent to complete. It’s more like a chatbot than a coding environment. Press Enter, the AI ​​agent gets to work without the developer needing to write a single line of code. In a sidebar on the left, developers can view and manage all the AI ​​agents they run in Cursor.

What’s unique about Cursor 3, compared to desktop apps for Claude Code and Codex, is that it integrates an agent-first product with Cursor’s AI-powered development environment. In a demo, Alexi Robbins, Cursor 3’s other co-lead of engineering, showed WIRED how users can invite an agent in the cloud to launch a feature and then review the locally generated code on their computer.

Nelle and Robbins say it doesn’t matter which interface developers spend their time on: they just want people to use Cursor.

Compete with AI Labs

I visited Cursor’s office in San Francisco’s North Beach neighborhood last week. The startup is reportedly raising new capital at a valuation of $50 billion, nearly double what it was valued in a funding round last fall, and has expanded into a former movie theater. Cursor employees used to throw their shoes in a pile by the door upon entering, but now there is a row of large shoe racks, signaling how the company is growing.

Yet Cursor still looks like a startup. Employees tell me that’s part of the appeal of working there; the company can ship quickly and doesn’t feel too corporate. But as it finds itself racing to catch Anthropic and OpenAI in the agent coding race, this demolition might not be enough. This battle, the one to create the best AI coding agent, is perhaps Cursor’s most capital-intensive chapter yet.