Using Open-Deep-Research with Inference Endpoints

[waawaa]

I tried to run Open-Deep-Research with an inference endpoint in a way that's possible to deploy for example deepseek-r1:670B but it seems no possible to use it.
I tried using Ollama / LiteLLM, but no option to configure it correctly, is it possible to use it, or the model class won't work with it?

[boxabirds]

Use Windsurf or Cursor to tweak the run.py file. It's this way I could get it going with Claude for instance.