Feature request: Filter and bulk-clean Library files with no remaining chat references

I use ChatGPT heavily, and my Library has grown to more than 3,000 files.

A large portion of these are screenshots, images, PDFs, text files, and other documents that I uploaded only as temporary context for individual conversations.

The current Library behavior creates a serious cleanup problem: when those conversations are later deleted, the uploaded files can remain saved in Library. Over time, this can leave hundreds or thousands of files that no longer have any useful conversational context.

At that scale, the main problem is not simply storage. It becomes extremely difficult to determine which files are still referenced by something useful and which files are effectively orphaned.

What makes this especially frustrating is that ChatGPT already appears to know the relationship between a file and its originating conversation in some cases.

For example, for some image files, the Library UI can show:

  • View original chat
  • This chat was deleted

So the system appears to have at least some metadata about whether a file’s originating conversation still exists.

What I would really like is for that relationship to become usable as part of Library cleanup.

Requested improvements

  1. Add a Library filter for reference status, for example:

    • No remaining chat references
    • Original chat deleted
    • Original chat unavailable
    • Still referenced by an existing chat
    • Still referenced by a project
    • No associated chat
  2. Allow users to bulk-select, review, and delete files with no remaining references.

    This is the most important part for me. I want to be able to find files that are no longer connected to any existing conversation or project, review them together, and delete them in bulk.

  3. Add a dedicated Storage cleanup view, showing categories such as:

    • Files with no remaining chat or project references
    • Files from deleted chats
    • Files whose original chat is unavailable
    • Files still referenced by existing chats or projects

    It would also be useful to sort these by:

    • File size
    • Upload date / age
  4. Show Original chat / Related chats consistently across file types.

    With some images, I can see View original chat or This chat was deleted.

    But with other uploaded file types, such as PDFs, TXT files, and other documents, the same relationship may not be shown at all, even when the file was uploaded directly into an existing conversation.

    This consistency would be useful, but for me it is secondary to having proper filtering and cleanup tools.

  5. When deleting a conversation, optionally offer:

    “Also delete files that are only used by this conversation.”

    This should be conservative. If a file is reused in another chat, referenced by a project, shared, or intentionally retained or organized in Library, it should be preserved.

    A review screen before deletion would make this safer by showing exactly which files would be removed.

Why this matters

Library is useful as persistent storage, but not every upload is intended to become a permanent asset.

Screenshots, one-off PDFs, text files, and temporary documents are often uploaded simply to explain something in a conversation. Once that conversation is deleted, those files can become digital clutter.

At small scale, this is manageable. At thousands of files, it becomes almost impossible to clean up reliably because the context that explains why each file exists is gone.

The ideal solution would therefore be reference-aware Library cleanup: let users see which files are still referenced by existing chats or projects, identify the ones with no remaining references, bulk-select them, review them, and delete them safely.

Since ChatGPT already exposes states such as View original chat and This chat was deleted in parts of the Library UI, exposing that relationship as searchable and filterable Library metadata would make large-scale cleanup much more practical.