Local llm

[D3m0cr1tusGR]

It would be great if it could connect to a local AI

[xaitax]

Do you mean like running Ollama locally and then talking to it?
https://github.com/ollama/ollama

[D3m0cr1tusGR]

Do you mean like running Ollama locally and then talking to it?
https://github.com/ollama/ollama

Yes, I just saw that there's a guide on how to use AppFlowy with Ollama on desktop. Thank you! Also, it will be cool to have this option in the self host

[hobgoblinsmaster]

I'm able to use a local Ollama (also tested with llama.cpp) by exporting OPENAI_BASE_URL="http://127.0.0.1:11434/v1" before launching sploitscan.
Unfortunately, I needed to modify the model name in /usr/lib/python3/dist-packages/sploitscan/sploitscan.py as it is hard coded.
2 vars should be added to the configuration file :
openai_base_url ⇒ to be use with any openai compatible local or distant llm.
openai_model ⇒ chose the preferred model