## Summary The desktop app's embedded browser pane repeatedly freezes/goes blank and has to be retried, caused by the app's GPU (graphics) sub-process crashing and restarting. This is a recurring pattern across multiple sessions, not a one-off. ## Environment - App version: `1.37937.3` (per `config.json`), with `1.40609.0` staged as a pending update at the time of filing - OS: Windows 11 Pro, build 26200 - GPU: AMD Radeon 780M (integrated), on a Ryzen 7 8745HS - GPU driver: AMD Adrenalin 26.7.1

Embedded browser pane repeatedly crashes/freezes due to GPU process crashes (AMD Radeon 780M)

Status Open

Maintainer reply None cached

Activity 0 comments · opened Aug 28, 2026

Summary

The desktop app's embedded browser pane repeatedly freezes/goes blank and has to be retried, caused by the app's GPU (graphics) sub-process crashing and restarting. This is a recurring pattern across multiple sessions, not a one-off.

Environment

App version: 1.37937.3 (per config.json), with 1.40609.0 staged as a pending update at the time of filing
OS: Windows 11 Pro, build 26200
GPU: AMD Radeon 780M (integrated), on a Ryzen 7 8745HS
GPU driver: AMD Adrenalin 26.7.1 (released 2026-07-24) — confirmed up to date via AMD Software at time of filing
32 GB RAM

Evidence

main.log (%LOCALAPPDATA%\Claude\logs\main.log) contains repeated entries like:

2026-08-24 11:01:07 [info] GPU process gone: {
  reason: 'crashed',
2026-08-25 12:10:20 [info] GPU process gone: {
  reason: 'crashed',
2026-08-25 12:13:54 [info] GPU process gone: {
  reason: 'crashed',
2026-08-27 18:31:39 [info] GPU process gone: {
  reason: 'crashed',
2026-08-27 18:33:40 [info] GPU process gone: {
  reason: 'crashed',
2026-08-27 18:36:11 [info] GPU process gone: {
  reason: 'crashed',

The three events on 2026-08-27 happened within a 5-minute window while repeatedly loading a JS-heavy single-page app (Cloudflare's dashboard, dash.cloudflare.com) in the embedded Browser pane — each one coincided with the pane visibly freezing/going blank, requiring a retry to recover.

Because the GPU sub-process crashes and respawns rather than the main app process dying, these incidents do not show up in Windows' Application Error/Hang event log — only in the app's own main.log. That made this hard to diagnose from the user side.

Suspected cause

This matches a known pattern of GPU-process instability between AMD Radeon 780M-class integrated GPUs and Chromium/Electron's GPU compositor, seen in other unrelated Electron apps, e.g.:

https://github.com/CapSoftware/cap/issues/1954 (crash on launch, atio6axx.dll, AMD 780M)
https://pcforum.amd.com/s/question/0D5Pd00000v1YtwKAE/amd-780m-gpu-driver-timeoutscrashing-when-watching-videos-not-in-games

It appears to be a hardware-acceleration/compositor compatibility gap rather than something an outdated driver would explain — the driver here is current.

Request

Could the app expose a user-facing "disable hardware acceleration" toggle (or document an existing launch flag, if one already exists internally)? I could not find any such setting in the app's Settings UI or in its config files (checked config.json for any GPU/hardware-acceleration key — none present). Disabling GPU-accelerated rendering is the standard mitigation for this class of Electron/Chromium bug and would let affected users trade a little rendering smoothness for stability without needing a fix to the underlying Chromium/AMD driver interaction.

Happy to provide more log excerpts if useful.

View original on GitHub ↗