Feature Request: Make Project Memory Transparent, Searchable, and User-Controlled

Feature Request: Make Project Memory Transparent, Searchable, and User-Controlled

I use ChatGPT Projects extensively for long-term work. Over time, I have realized that Projects are becoming much more than simple folders for organizing chats. For many users, they are becoming long-term collaborative workspaces.

A single project may contain many conversations over weeks or months, covering troubleshooting, research, decisions, failed attempts, conclusions, documents, and evolving plans .

This creates a new need: not only project-scoped retrieval, but transparent and user-controllable project continuit y.

Current is sue

When working inside a Project, users often do not know whether ChatGPT is actually using previous conversations from that Proj ect.

Sometimes, important context already exists in older project conversations, but the user still has to manually summarize or paste it a gain.

This creates uncertainty:

• Did ChatGPT search the current Project?
• Did it only use the current conversation?
• Did it use uploaded files but not past conversations?
• Did it miss an important previous decision?
• Is it relying on global memory instead of project-specific knowledge?

For long-term projects, this uncertainty becomes a real limitation.

The m issing piece

Project knowledge is not only stored in upload ed documents.

It is also stored in the conversatio ns themselves.

For example, a long-term proje ct may contain:

• decisions that were made months ago,
• assumptions that were tes ted and rejected,
• troubleshooting steps that worked,
• troubleshooting steps that failed,
• technical configurations,
• writing direction,
• research conclusions,
• user preferences specifi c to that project,
• explanations built over many conversations.

This is not just “chat history.”

It is accumulat ed project knowledge.

Proposed improvements

I would like to suggest making Project Memory and Project Retrieval more vi s ible and controllabl e.

Possible features:

1. “Search this Project before answering”

Add a button or command that lets the user explicitly ask ChatGPT to search the current P roject bef ore responding.

Example:

Search th is Project before answering.

This would be very useful when the user knows the answer may already exist somewhere in pr e vious project conversat ions.

2. Show what was used

After answering, ChatGPT could indicate what kind of c ontext was used.

For example:

Used context from: current conversation, 3 previous project conversations, 2 uploaded files.

This would build trust and help users understand whether the answer was properly grounded in the Project.

3. Let users inspect retrieved context

Users could optionally expand a small section showing which previous convers ations or documents were referenced.

This would make the system more transparent witho u t cluttering the default experie nce.

4. Add project retrieval modes

Eac h Project could have a setting such as:

• Current conversation only
• Current conversation + uploaded files
• Current conversation + previous project conversations
• Full Project context when relevant
• Ask each time

This would give users more control depending on the se nsitivity and complexity of the project.

5. Dist inguish global memory from project memory

Global mem ory should store general user preferences.

Project memo ry should store project-specific knowledge.

These two layers should remain clearly separated, so unr e lated projects do not influence each other.

Why this matters

For s i mple conversations, this may not matter muc h.

But for long-term work, it matters a lot.

Examples include:

• software development,
• technical troubleshooting,
• research,
• engineering,
• legal or administrative work,
• writing projects,
• education,
• investment research,
• creative worldbuilding,
• personal knowledge management.

In these case s , users are not just asking isolated que s tions.

They are b uilding knowledge over time.

The broader vision

Projects should not only organize conversations.

They should help preserve and reuse the reasoning, decisi ons, and knowledge built inside those conversations.

In other words, Projects could evolve f rom folders into persistent collaborative workspaces.

Project Memory should not feel mysterious or passive.

It should be:

• visible,
• searchable,
• controllable,
• project-specific,
• and trustworthy.

I believe this would make ChatGPT much more powerful for users w ho rely on it as a long-term research and work partner.

As more people use ChatGPT for months or years instead of isolated conversations, Projects are evolving into long-term collaborative workspaces. Making Project Knowledge transparent and controllable would help bu il d the trust required for that kind of colla b ora tion.

Thank you for considering this suggestion.