We've spent the past year improving the Next.js agent experience. Along the way, we built and sunset an in-browser agent, shipped MCP integration, and learned that the real answer to better agent support is thinking from the agent's perspective.
Earlier this summer, we were working on improving the Next.js devtools when we noticed a pattern. Developers would see an error in the browser, copy the details, paste them into an AI editor, and ask the agent to fix it.
The problem was that agents can't see the browser. Runtime errors, client-side warnings, and rendered components are all invisible to them. When a user says "fix the error," the agent doesn't know what error they mean.
This is an excerpt. The full article is published by Next.js Blog.
Continue reading at Next.js Blog