Open pirateben820 opened 9 months ago
Upvote.
Simple integration for advanced users (could be up and running tomorrow): Allow users to specify the API endpoint. I already have local llms running on my own fastapi that's 100% compatible with the OpenAI API. Advanced users tweak their api on their end.
Advanced integration for newbs: File picker for quants, or Cursor-specified quant models living on a Cursor HF that are pre-set in the software.
Would love to see the simple integration soon @truell20 ! Thanks for everything you all do, Michael + Cursor team!
Interesting implementation for Copilot for Xcode: https://github.com/intitni/CopilotForXcode
Yeah, been wanting that for a while now as well
Upvote here :) I want to use my local mistral model in Cursor :) I want to be able to work offline. I want to use open source AI model. I want to cut down cost of OpenAI model. Please make this happen :)
They recently added the ability to override the API key, but LM Studio endpoint still does not work.
Specifically, this error shows up:
On LM Studio side, this shows up:
So cursor is sending an empty request to OpenAI to check if the key works, but LM Studio did not account for this use case in the API endpoint. So either LM Studio will have to fix this, or Cursor will need to ignore this error for it to work correctly.
Is your feature request related to a problem? Please describe. not a problem but a feature option that i think would be easy to integrate also other then using your openai api key have a pointer to a serves that sends out data via server
Describe the solution you'd like by adding local host or network servers to use there ai people can play with free opensource modes and what not here is a video of lm studio server https://youtu.be/yBI1nPep72Q?t=619 Additional context by doing this it is a free backend only the user has to deal with and teams can make and pick the ai they want