{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreiaff6aajm6hilp3j7hyvjcahfyqusv4kccvwi46k7o2ohcftmna6e",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mn44ahzh3ew2"
  },
  "path": "/t/feature-request-compact-long-chatgpt-conversations-into-a-clean-new-thread/1382084#post_3",
  "publishedAt": "2026-05-30T20:39:19.000Z",
  "site": "https://community.openai.com",
  "textContent": "One important requirement: relevant files and artifacts should be carried over as well.\n\nLong-running ChatGPT workflows are often not just text conversations. They may include uploaded PDFs, screenshots, spreadsheets, code files, generated documents, images, canvases, or intermediate outputs.\n\nIf a compacted new chat only carries over the summary but loses the relevant files/artifacts, the user has to manually re-upload or reconstruct the working context, which weakens the value of the feature.\n\nSuggested behavior:\n\n  * identify files/artifacts still relevant to the active source of truth\n  * let the user choose which ones to carry over\n  * exclude files tied only to discarded or invalidated directions\n  * preserve file names and provenance\n  * optionally include a note explaining why each file was carried over\n  * keep a link back to the original conversation\n\n\n\nThe ideal flow would be:\n\nCompact into new chat\n→ generate editable handoff summary\n→ show active / superseded / invalidated context\n→ show relevant files and artifacts to carry over\n→ user confirms\n→ create clean new thread with both the summary and selected files attached\n\nThis matters because the new thread should not just preserve a text summary. It should preserve the usable working state.",
  "title": "Feature request: Compact long ChatGPT conversations into a clean new thread"
}