how to use this extension if i already got an openai compatible api?

it keepss asking me to install local server on linux and in order to install it id have to fully download a useless model which i already have hosted and running on remote server which im connecting to

where am i supposed to write questions?

https://github.com/ggml-org/llama.vscode/wiki/Llama-agent

clearly some broken piece of trash

[igardev]

If you want to use remote server, just say "No" to this question. You don't need to install a local model.
Just note that for the code completion features (Compl Model) you need to use llama.cpp server with a model, which supports FIM. For the other models just openai compatible server is OK.

How to setup a remote server:

• Click the appropriate "More" button for the model (or Ctrl+Shift+M -> * models -> Add external * model)
• Enter model name
• Enter the URL to the server (llama.cpp or openrouter.com, etc.) - for example http://localhost:8011
• If needed, enter the name of the LLM model
• Answer the remaining questions
• Use the "Select" button to select the model, which was just configured and start using it

Note that the API key vlaue need to be entered only when the model is selected for the first time.