https://promptfoo.dev/ logo
Question: Is it possible to pass Pinecone embeddin...
# ❓|questions
a
Hey there, my LLM application uses Pinecone as RAG context to generate outputs. I was wondering if there's anyway to pass Pinecone embeddings as context in promptfoo so that the score I get gives a more accurate representation of my app's performance as opposed to directly putting text from my documents.
i
Hi @anoushkaramesh, you would have to pass the documents into the context because promptfoo is not hooked up directly to your retrieval system. A [custom script provider](https://promptfoo.dev/docs/providers/custom-script) is a good way to do this - you can call out to your existing embedding retrieval logic. You may also be interested in the [RAG evaluation guide](https://promptfoo.dev/docs/guides/evaluate-rag)
a
Hey thanks! This helps