Claude Code can now test its own software
Anthropic just added computer-use capabilities directly to Claude Code. The CLI agent can now open apps, click through user interfaces, and automatically test the software it just built before handing it back to you. The feature is rolling out as a research preview for Pro and Max subscribers.
Coding agents are great at churning out software, but actually opening the app and clicking around to see if it works has always fallen on the developer. Until now.
Anthropic just rolled computer use capabilities directly into Claude Code. The update is currently available as a research preview for Pro and Max subscribers.
Closing the loop in the CLI
Everything happens straight from your terminal.
Claude can now spin up the applications it just built, click through the actual user interface, and run its own end-to-end tests without leaving the command line.
Basically, it takes the tedious manual step of post-build UI verification off the developer’s plate.
Why it matters
We think this is a huge step toward true autonomy for coding agents.
The bottleneck with AI coders hasn't just been generating syntax; it has been the verification loop. Giving Claude Code hands to actually interact with what it built turns it from an autocomplete engine into a full-cycle tester.
That said, handing desktop-level screen control to a CLI agent carries obvious risks around reliability and accidental clicks. It looks like that is exactly why Anthropic is keeping this in a gated research preview for paid tiers first.
Sources (1)
- x.com(opens in a new tab)Originalx.com