Dear OpenAI Team,
I am a ChatGPT Plus user, and I would like to submit a product feature request based on an issue I recently encountered while using ChatGPT for a long-term project.
I have been working with ChatGPT on the same project for a long period of time. During this process, the conversation accumulated a large amount of important context, including project architecture, implementation decisions, completed work, troubleshooting history, files, technical constraints, and planned next steps.
Eventually, ChatGPT informed me that the conversation had reached its length limit and that I needed to continue in a new chat.
I understand that there may be technical limits related to conversation length, context management, storage, or model performance. My concern is not necessarily that conversations should be unlimited.
The main problem is what happens after a conversation reaches its limit.
When I moved to a new conversation, the new ChatGPT instance did not automatically understand the full state of the project. I had to spend a significant amount of time asking the previous conversation to summarize the project, preparing documents, transferring information, and then helping the new conversation understand our previous working method, technical decisions, current progress, and unresolved issues.
For users who use ChatGPT for long-term engineering, research, software development, study, or personal projects, this creates a major continuity problem.
I would therefore like to suggest a Conversation Handoff / Project Continuity feature.
Ideally, when a conversation is approaching or reaches its length limit, ChatGPT could provide options such as:
- One-click export of the current conversation
Allow the user to export the specific conversation directly, rather than requesting a full account data export.
Possible formats could include Markdown, JSON, PDF, or a ChatGPT-native conversation package.
- One-click structured project summary
ChatGPT could automatically generate a structured handoff document containing information such as:
-
Project objective
-
Current architecture or design
-
Important decisions already made
-
Completed work
-
Current implementation status
-
Files and resources being used
-
Technical environment
-
User requirements and constraints
-
Problems already investigated
-
Known failures and solutions attempted
-
Unresolved issues
-
Next recommended actions
-
Important context that the next ChatGPT conversation should know
- “Continue in a new chat” button
Instead of simply telling the user that the conversation has reached its limit, ChatGPT could offer a button such as:
Continue this project in a new chat
The new conversation could automatically receive the handoff summary and relevant project context.
From the user’s perspective, it would feel like continuing the same project rather than starting over with a new AI.
- Advance warning before the conversation limit
It would also be helpful if ChatGPT warned users before the conversation reaches its maximum usable length.
For example:
“This conversation is approaching its context limit. Would you like ChatGPT to prepare a project handoff summary?”
This would allow the original conversation, which still understands the project best, to prepare the handoff before the limit is reached.
- Persistent project-level context
For projects that contain multiple conversations, it would be extremely useful if important project information could exist independently of any individual chat.
For example, a Project could maintain a persistent project state containing:
-
project goals
-
technical architecture
-
important decisions
-
files
-
progress
-
current tasks
-
project-specific instructions
Individual conversations could then reference this shared project state.
I am aware that ChatGPT currently provides account-level data export, and that exported conversation data can be uploaded into another conversation as reference. However, this workflow is different from an integrated project handoff system and is not practical when a user simply needs to continue an active project immediately.
For long-term ChatGPT users, especially Plus and Pro users who use ChatGPT as a serious productivity and project-development tool, I believe this feature could significantly improve reliability and continuity.
In my case, the project itself was not interrupted because of a technical failure in the project. It was interrupted because the conversation containing the project’s working context reached its limit.
I believe ChatGPT should be able to preserve the state of the work, even when it cannot preserve an unlimited conversation.
Thank you for considering this feature request. I would be very interested in seeing a project handoff or conversation continuity system added to ChatGPT in the future.
Best regards,
A ChatGPT Plus user