Well Needed Feature For LIBRARY

ChatGPT Feature requests
,
1

Feature Request: ChatGPT Librarian

I’d like to propose a dedicated Librarian experience for ChatGPT: a persistent knowledge-management agent responsible for organizing, indexing, versioning, retrieving, and maintaining a user’s ChatGPT Library across long-term use.

The problem I’ve encountered while using ChatGPT for substantial, ongoing development projects is that Projects and Work can accumulate enormous amounts of valuable information. As histories grow, older specifications, superseded decisions, current requirements, files, and project artifacts become increasingly difficult for both the user and AI to distinguish reliably.

A Librarian could maintain a machine-readable knowledge index across the user’s Library, identifying relationships such as Current, Approved, Draft, Superseded, Historical, Experimental, and Unresolved, while preserving the underlying source material.

This would benefit not only human organization but AI retrieval: future ChatGPT sessions could consult the Librarian’s structured index before retrieving historical source material, reducing continuity errors and outdated instructions being mistaken for current ones.

Projects organize work. Memory understands the user. Librarian organizes knowledge.

I have developed a more detailed product proposal describing the workflow, information hierarchy, versioning model, user controls, and potential integration with Projects and Work, and I’d be happy to provide it if useful.

2

Hi @the.webguy,

Thank you for sharing this thoughtful proposal. A structured knowledge index could be valuable for users working on projects that span many conversations.

One practical improvement that could complement your idea is enhancing the “Add from Space” file picker. As saved files accumulate, choosing the right document becomes harder, especially when filenames are similar or several versions exist.

The attached mockup illustrates a few improvements:

  • Search and filtering by file type, project, or folder, helping users narrow down results quickly.

  • Sorting by name, upload date, last modified date, or size, making relevant files easier to locate.

  • Full filenames and clear file details, helping users distinguish similar documents.

  • A preview before attachment, so users can check the content without leaving the conversation.

  • Multiple-file selection with a count and total size, making it easier to bring related sources into a new chat.

  • An option to delete unwanted saved files directly from the picker, with explicit confirmation that this removes them from the Library. This would let users clear unnecessary files without interrupting their workflow.

Another useful addition would be a contextual prompt when an uploaded file matches content already saved in the Library, offering to reuse the existing file. I only recently discovered “Add Space files,” so this could help users discover the feature while avoiding unnecessary repeat uploads.

For the Librarian itself, an index over existing content, supported by stable file references, could improve retrieval while preserving links from earlier conversations.

These changes would help users spend less time finding and checking files, attach the right sources with greater confidence, and continue their work more smoothly.
@PaulBellow, I’d be interested to hear your thoughts on whether these file-picker improvements would complement the Librarian proposal.

image
image1536×1024 159 KB

image
image1371×994 21.1 KB

3

Looks like a decent mock-up of the idea. Thanks for taking the time! :slight_smile:

4

I’m just the idea guy.lol Sounds like you’re the developer person. I could imagine a plethora of useful features added to make the Librarian a native plugin.

The Add From Space idea seems fantastic, however, the ideal Librarian would do all that for you-removing the need to select, choose, navigate, or waste precious time organizing.

Awesome work though!!