[BUG] New In-app browser causes system instability/memory pressure on 8GB Macs

Open 💬 0 comments Opened Jul 16, 2026 by joylarkin

Preflight Checklist

  • [x] I have searched existing issues and this hasn't been reported yet
  • [x] This is a single bug report (please file separate reports for different bugs)
  • [x] I am using the latest version of Claude Code

What's Wrong?

Environment:

macOS Tahoe (26.5.2, build 25F84)
MacBook Air, Apple M2, 8GB RAM
Claude Code desktop app (macOS)
Description:

For several days this week, my system was unreliable (sluggish, unresponsive) with no obvious cause. After eliminating other suspected problem areas, I traced the instability to the in-app browser feature in the Claude Code macOS desktop app. After disabling it in Settings, my system reliability improved noticeably.

On resource-constrained machines (8GB RAM - yes I know it's a ridiculously underspec'd machine from a previous startup), the in-app browser appears to consume enough memory/resources to cause broader system instability, even when not actively in focus. The instability behavior happened when I merely opened the Claude Code desktop macOS app.

What Should Happen?

The Claude Code macOS app should be able to be opened and used on my machine without causing grave memory issues.

Error Messages/Logs

Steps to Reproduce

Steps to reproduce:

On an 8GB Mac, enable the in-app browser in Claude Code desktop app settings
Use the app normally for a few days
Observe general system slowdown/unresponsiveness
Disable the in-app browser in Settings
System reliability improves

Claude Model

_No response_

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.150

Platform

Other

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

Suggestion:

  • Consider a lower-memory-footprint mode for the in-app browser, or a warning/prompt for users on systems with limited RAM
  • Would also be helpful if the app surfaced resource usage per-feature so users can self-diagnose issues like this faster

Thank you for your help -- and yes I actually do prefer using the desktop app over the terminal.

View original on GitHub ↗