openai / openai/codex

Codex Desktop: missing model provider error toast repeatedly flashes and flickers chat UI

Open
#46,602 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

app bug custom-model session windows-os
Dominant language
Rust
Stars
125k
Forks
19.4k
PR merge metrics
PR metrics pending

Description

Environment
  • Codex Desktop on Windows 11
  • App version: 26.915.31029 (reported in local configuration; not independently checked in About dialog)
Summary

When opening an older conversation whose session metadata references a model provider that is no longer configured (custom-gateway), Codex correctly refuses to resume it, but the error notification repeatedly appears and disappears roughly once per second. The entire chat UI flickers as the notification comes and goes, making the message nearly impossible to read.

This issue is about error presentation, not a request to load a conversation with an unavailable provider.

Steps to reproduce
  1. Have an existing conversation whose session metadata specifies model_provider: "custom-gateway", without a corresponding provider definition in the current config.toml.
  2. Open Codex Desktop and select that conversation from the chat list.
  3. Observe the error notification at the top of the conversation.
Actual behavior
  • The red notification flashes briefly and disappears, then reappears repeatedly, approximately once per second.
  • The chat interface visibly flickers with it.
  • The notification is too brief to read normally; a screen recording had to be paused to capture the text.
  • The conversation does not resume, as expected given the missing provider.
Expected behavior

Show the configuration error once in a stable, readable notification (ideally persistent until dismissed, with an option to copy its text). The conversation may remain unavailable until the provider configuration is fixed.

Error text

ChatGPT can't load config.toml, so this thread can't resume.

Fix config.toml: Model provider custom-gateway not found. After saving the file, reopen the thread.

Additional observations
  • New conversations work normally.
  • A newer conversation with openai as its model provider opens normally.
  • Five older local session files had model_provider: "custom-gateway" in their first line; one inspected session originated from a JetBrains Rider integration.
  • Restarting the desktop app and signing out/in did not change the behavior.
  • No session files were edited as part of the investigation.

A screenshot of the notification is available; no private session files or configuration containing local paths are attached.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reproducing the missing-provider conversation on Codex Desktop for Windows with model_provider set to custom-gateway and no matching config.toml entry. Trace why the error notification is repeatedly mounted and dismissed, then verify that the configuration error remains readable and stable while the conversation stays unavailable.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
desktop
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
58/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.