Issue Station
Sign In
Tools hereToolbox

codex - 💡(How to fix) Fix [Accessibility] NVDA focuses on previous messages instead of latest response after recent update [1 comments, 2 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openai/codex#17530•Fetched 2026-04-12 13:26:56
View on GitHub
Comments
1
Participants
2
Timeline
5
Reactions
0
Author
ademalgulademalgul
Participants
ademalgulademalgulgithub-actions[bot]github-actions[bot]
Timeline (top)
labeled ×3commented ×1subscribed ×1
RAW_BUFFERClick to expand / collapse

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

26.406.31014

What subscription do you have?

Enterprise

What platform is your computer?

Microsoft Windows 10 64-bit (x64)

What issue are you seeing?

Hello,

I am a blind user using Codex Desktop App on Windows 10 64-bit with NVDA screen reader.

After the recent updates, a significant accessibility issue has appeared between Codex and NVDA. When Codex responds to my messages, NVDA does not focus on the latest (newest) response. Instead, the focus jumps to previous messages or earlier parts of the thread. This makes it very difficult and time-consuming to read the new answer every time.

In contrast, the regular ChatGPT web interface works very well with NVDA. In ChatGPT:

  • My messages and the assistant’s responses are clearly separated as distinct message blocks.
  • Even in very long conversations with hundreds of messages, NVDA can easily navigate between them.

I believe implementing a similar clear message structure (like in ChatGPT) in Codex would solve this focus problem completely.

Could you please look into this accessibility issue? It would make Codex much more usable for screen reader users.

Thank you.

What steps can reproduce the bug?

Steps to reproduce the bug:

  1. Open Codex Desktop App with NVDA screen reader.
  2. Open any existing thread (especially a longer one).
  3. Type a message and press Enter to send it.
  4. Wait for Codex to generate and complete the response.
  5. NVDA does not move focus to the newest response. Instead, it stays on or jumps to previous/older messages in the thread.
  6. I have to manually press NVDA + End or use arrow keys to reach the latest message every single time.

This happens consistently after the recent update. Expected behavior:

After I send a message and Codex finishes responding, NVDA should automatically focus on the latest (newest) message so I can read the response immediately without manual navigation.

Similar to how ChatGPT web interface works with NVDA — each message is clearly separated and focus moves naturally to the new response.

What is the expected behavior?

No response

Additional information

No response

extent analysis

TL;DR

Implementing a clear message structure in Codex, similar to ChatGPT, may resolve the NVDA focus issue.

Guidance

  • Verify that the issue is specific to the Codex Desktop App by comparing its behavior with the ChatGPT web interface, which works as expected with NVDA.
  • Check if there are any existing accessibility settings or updates in Codex that could be contributing to the focus issue.
  • Consider testing with other screen readers to determine if the issue is specific to NVDA or a more general accessibility problem.
  • Review the HTML structure of the conversation threads in Codex to identify potential differences with ChatGPT that could be causing the focus issue.

Example

No specific code example can be provided without more information about the Codex App's internal implementation.

Notes

The issue seems to be related to the way Codex structures its conversation threads, which may not be providing the necessary accessibility features for NVDA to focus on the latest response.

Recommendation

Apply workaround: Implement a clear message structure in Codex, similar to ChatGPT, to improve accessibility with NVDA. This is recommended because the ChatGPT web interface already demonstrates a working solution with NVDA, and adapting this structure could resolve the focus issue in Codex.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#network issue#logging issue#authentication issue#prompt issue#agent setup

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING