{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreic6sedwyym4sljvlt3euzsxa4l2wnsvb6q7chuceym2lq4mn6jxme",
"uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mkph6rw2fqz2"
},
"path": "/t/compacting-error-problem-with-codex-app/1379941#post_3",
"publishedAt": "2026-04-30T09:43:46.000Z",
"site": "https://community.openai.com",
"textContent": "I also have been hitting this for months and months, it keeps breaking my workflow too. 20% context remaining usually and /compact just kill the session.\n\n■ Error running remote compact task: stream disconnected before completion: error sending request for url … backend-api/codex/responses/compact)\n\n■ Error running remote compact task: stream disconnected before completion: error sending request for url … backend-api/codex/responses/compact)\n\nIt’s really breaking my workflow. Each time I have to start a new session and tell the model to read the tail of the previous convo history in the .codex folder and resume. I can barely get any session that compacts more than 2 or 3 times. It’s more frequent recently. I use multiple providers, and some of them can run uninterrupted for 5 hours. This is a real issue that needs solving. Perhaps allow setting the auto compact threshold like at 40%, at least the session will keep working instead of permanently breaking every 15-30 minutes.",
"title": "Compacting error - problem with Codex App"
}