[macOS] “More details” Quick Chat turns white when sending a large text input

What version of the Codex App are you using (From “About Codex” dialog)?

26.803.81509 (build 6415)

What subscription do you have?

Not provided

What platform is your computer?

macOS 26.5.2 (build 25F84), Apple Silicon (arm64)

What issue are you seeing?

In the Codex desktop app, selecting text and choosing More details opens the ChatGPT Quick Chat surface as expected. However, when a large amount of text is entered or pasted into that Quick Chat and sent, the surface suddenly becomes entirely white and enters an error state. The Quick Chat can no longer be used normally.

This appears to occur in the ChatGPT handoff/Quick Chat opened from More details, rather than after adding that conversation back into the Codex task.

What steps can reproduce the bug?

1. Open a Codex task in the macOS desktop app.
2. Select text in a message and choose More details.
3. In the ChatGPT Quick Chat that opens, enter or paste a large amount of text.
4. Send the message.
5. Observe that the Quick Chat surface becomes a blank white screen/error state.

The exact character threshold has not yet been isolated, but the failure is associated with a large text input.

What is the expected behavior?

The Quick Chat should accept and render a large text input normally, or provide a clear, recoverable validation/error message if an input limit is exceeded. It should not become a blank white screen.

Additional information

Potentially related but distinct from #31966 and #32033: those reports describe Codex freezing when a More details conversation is added back to the task. This report concerns the Quick Chat itself becoming blank immediately after sending a large text input, before that return-to-task flow.

A screenshot/screen recording can be provided if needed.

Start by reproducing the issue through the macOS desktop app's More details entry point: open Quick Chat, send progressively larger text inputs, and observe when it becomes blank. Compare the behavior with the expected recoverable rendering or validation outcome, and verify that the Quick Chat remains usable after sending large input.