Dear OpenAI,
I’m running into a structural limitation with ChatGPT exports that becomes apparent once you use ChatGPT for multiple long-running subjects.
My usage is topical, not chronological. Each subject (technical projects, planning, research, etc.) develops across multiple threads over time.
To maintain continuity, I currently upload collections of subject-specific threads as text files at the start of each new thread in that subject. This works, but only after manually extracting the relevant parts from the full export.
The problem is that the export format is purely chronological, while the work itself is subject-based:
-
The ChatGPT export is chronological, not topical
-
My work is topical, not chronological
-
As subjects multiplied, the “export everything” approach stopped scaling
When starting a new thread in a subject, I know the relevant context exists — but it’s buried among unrelated threads from other subjects. This forces unnecessary manual cleanup just to preserve continuity.
What would be extremely useful:
-
The ability to selectively export conversations (by thread, tag, or subject)
-
Or export a subset of threads rather than the entire account history
-
Even a basic “export selected conversations” option would significantly improve usability for long-term, project-based use
This isn’t about exceeding context limits — it’s about maintaining continuity in complex work without dragging unrelated history forward.
I realize this may go beyond the original casual-chat use case, but for users treating ChatGPT as a long-term workspace, selective export would be a major quality-of-life improvement.
Thank you 