should work with OpenAI-compatible self-hosted endpoints

[7flash]

Currently I can switch to langchain when using OpenAI

- import openai
+ from langfuse.openai import openai

But in case if I am hosting my own model with openai-compatible API using vllm or litellm, I would like still to benefit from langfuse observability

os.environ["OPENAI_ENDPOINT"] = ""

[Manouchehri]

If you're using LiteLLM, then this should already work. We're using it already. https://docs.litellm.ai/docs/proxy/logging#logging-proxy-inputoutput---langfuse

[marcklingen]

Closing this thread as this works already when changing the base url of the sdk.