Hi,
I’m considering using MPC integrations in Cursor to access internal documentation like Confluence. However, I’m concerned about data security.
If Privacy Mode is turned ON, does it completely prevent data from being sent to external servers, including AI models? Or is there still a risk that some information might be processed outside the local environment?
I’d appreciate any clarification on how Privacy Mode works and whether it ensures that no sensitive data is exposed.
Thanks in advance!
From what I understand, enabling Privacy Mode in Cursor ensures zero data retention, meaning your code and internal documentation won’t be stored or used for training by Cursor or third parties.
However, even with Privacy Mode on or when using your own API key, requests still go through Cursor’s backend, so some external processing still happens—likely to prevent access to their system prompts.
If you’re looking for a fully local experience with no external processing, you might want to check out Zed editor with Ollama, which allows running models entirely on your machine.
You can also review Cursor’s Privacy Policy and Security for more details.