I believe there is a bug in the factory.py when using ResponseMode.NO_TEXT. I have different deployments with different API keys in Azure. This doesn't work unless I add in the llm=llm parameter to the NoText constructor.

What is the cheapest / free way for a compsci student to use llamaindex to build a simple RAG based chatbot? Are there free APIs to create the vectors? Then to call some light version of something like chat?

Here's a bug in the 0.6 version. If I add the documents to the vector as follows:
index = GPTVectorStoreIndex([], service_context=service_context)
for i, doc_chunk in enumerate(chunks_so_far):
index.insert(doc_chunk)
then when it persists the storage_context , no doc_ids get written to the json file.
When I build it with the one line.
tree_index = GPTVectorStoreIndex.from_documents(chunks_so_far, service_context=service_context)
it works fine to save the index. I assume the first version should work.