Just wondering if local mode works when you use your own Openai key. From what I read in the documentation it works when you sign up for a Cursor plan, as they have an agreement with Openai. But it is not obvious if local mode restrictions are fulfilled when you use your own Openai key. Anyone know?
Local mode (i.e. none of your code is stored anywhere by us) works with your API key or a Cursor plan.
OpenAI will treat the data the same too (i.e. they will not train on your data but will retain it for 30 days for trust and safety monitoring).
Thanks for the clarification!
To clarify, in local mode dose any of our code pass through your systems?
Yes! Even with local mode, small chunks of code are sent up to our server to build a prompt that we send to OpenAI. No code is stored anywhere though; it ceases to exist after the life of the request.
The fact that this is called local mode makes me trust the company less. I work for a startup, too. We have cowboy marketing people, too. This is too far. Local mode implies that my code does not leave my device. Maybe you have plans to add support for local LLMs but that is not what exists now (and they don’t work well enough yet).
I recommend renaming it to secure mode. People will know to question what that means. If you leave it as local mode, you will piss off security managers that evaluate the software because their first impression/assumption will be violated. If it is called secure mode, they will be open minded about the implementation.
Hey @EatSushiNow! Appreciate the feedback. We decided to call it local mode because all of your code is stored locally, but if others feel similarly, we can look into finding a more appropriate name here.
I’m just one guy. It might be worth putting on a marketing survey in the future. Describe what it is in a paragraph and ask which name it should be given.