About using different LLMs such as DeepSeek

[mhammadpk]

@ShahabEJ

I tried setting up a different LLM, specifically DeepSeek, by changing the environment variable from OPENAI_API_KEY to DEEPSEEK_API_KEY and using DeepSeek API value. However, I'm encountering an error.

Could you please check if there are additional configurations needed, or updating the code as well? Let me know how to proceed.

Additionally, is it possible to structure the code in a plug-and-play manner, where simply changing the environment variable (e.g., OPENAI_API_KEY, DEEPSEEK_API_KEY, etc.) automatically adapts the entire setup to the corresponding LLM? This would allow seamless switching between different providers without requiring major code changes.

Would love your insights on this!

[ShahabEJ]

@mhammadpk, the error comes from the RAG functionality in the Knowledge Base as it also uses OpenAI text embedding model. If you don't use the Knowledge Base, you can keep the env variable for OpenAI but leave it blank so you don't get this error.