Feature request for product/service
Chat
Describe the request
Hello Cursor Team,
I couldn’t find a dedicated in-app feedback channel, so I’m posting this here as a feature request for the team.
First of all: great work on Cursor.
I really like the overall UX, editor integration, and how naturally AI assistance fits into day-to-day development workflows.
I’d like to share a constructive feature request that comes from working on a long-running, architecture-heavy project.
Context / Use Case
I am working on a complex, long-term system architecture with many interdependent Markdown specifications and code artifacts.
This kind of work relies heavily on extended reasoning continuity across sessions rather than short, isolated prompts.
While I have Cursor installed and my project fully loaded (including rules, markdown specs, etc.), I currently still rely on an external AI for a significant part of the work — simply because I need a persistent conversational context that I can actively control.
Inside Cursor, the lack of an explicit, configurable session context currently makes it difficult to:
- iteratively refine specifications
- audit earlier architectural decisions
- pause and resume work without losing reasoning continuity
This would also pair extremely well with Cursor Rules and long-lived project contexts.
Feature Request: User-Configurable Context Window
What would make a huge difference for advanced users like me would be a configurable, explicit context window, for example:
- A message-based, size-configurable context buffer (with token limits handled transparently)
- Clear visibility into what is currently in context
- FIFO-style aging of older messages
- The ability to pin important messages or documents so they are never dropped
- Transparent context management, including when and why older context is truncated
This would be especially valuable for:
- architecture & specification work
- long-lived refactors
- audits and design reviews
- non-trivial system design sessions
Why This Matters
Right now, the lack of such a mechanism is the main reason I haven’t fully moved my AI-assisted workflow into Cursor — despite actively wanting to.
With a configurable, persistent context:
- I would gladly shift more work into Cursor
- I would be very happy to subscribe to a paid / premium plan
- Cursor would become my primary environment for both code and architecture work
Closing
I fully understand that context management is technically and product-wise non-trivial.
Still, I believe this feature would unlock an entire class of long-running, architecture-heavy engineering workflows and significantly differentiate Cursor from simpler AI editor integrations.
Thanks for taking the time to read this, and thanks again for building a great tool.
Best regards,
Ilker