Most people don't have OPENAI API key.

[jeet-Abhi123]

As OPENAI API key is paid and most students can't afford to pay for learning purpose.
Can you please create the code using GROQ API inference providers. Because they provide access to open source models to some extent.
I have tried editing the code, edited the helper_functions.py, evaluate_rag.py but then there is a lot of dependency on Openai.
Your Github repo is a gem to understand RAG techniques in depth. So, can you please create one tutorial for Groq?
Link : Groq Inference Providers

[wangzhuo0601]

I found that the project had to use the official OPENAI_API_KEY and the transfer OPENAI_API_KEY was not used. It was difficult for me to obtain OPENAI_API_KEY, whether it could be expanded, such as GEMINI_API_KEY, etc

[NirDiamant]

the llm provider can be replaced easily with just one line of code. you can open a PR that shows some alternatives

[jeet-Abhi123]

Ok @NirDiamant thanks sir

[NirDiamant]

Valid point - not everyone has an OpenAI API key, and the repo should be accessible to as many developers as possible.

Several notebooks already support local models via Ollama (check the simple_rag and other foundational notebooks). The community has also contributed PRs for local-only pipelines (#104 adds Ollama support, #132 adds HuggingFace + FAISS).

If you're looking for a conceptual understanding of all 22 techniques without needing any API key at all, the book version covers every technique through visual diagrams and plain-English explanations: https://www.amazon.com/dp/B0D76734SZ

Thanks for the feedback - making the repo more accessible to developers without paid API keys is something we're actively working on.