do you need an api key with GPT 4.0 to

At a glance

The community members discuss whether an API key is required to use npx create llama with GPT 4.0. One community member states that an API key is not needed, but the LLM and embedding model will need to be changed if not using OpenAI. Other community members provide links to the Llama Index documentation for LLM and embedding models, and suggest trying the Python backend if the TypeScript backend does not support the desired non-OpenAI models. The community members seem to be exploring options to use Llama without an OpenAI API key.

Useful resources

ccore

do you need an api key with GPT 4.0 to use npx create llama?

6 comments

LLogan M

Dont dont NEED an API key, but you'll need to change the LLM and embedding model if you don't use openai

ccore

this https://ts.llamaindex.ai/modules/low_level/llm and this "https://ts.llamaindex.ai/modules/low_level/embedding" ?

ccore

and I can change it to one of theses? "https://docs.llamaindex.ai/en/stable/module_guides/models/llms/modules.html"

LLogan M

so that third link is the python library only

Assuming you used a typescript backend with create-llama, the llms are

anthropic
azure
openai
portkey
replicate

And embeddings are

openai

rip

LLogan M

So only the python backend supports non-openai embeddings right now it seems -- would love to see more embeddings added to the TS package

ccore

I'll try to get around the openai API key for now, maybe I'll learn more by doing it this way. I'll give the python backend a try

Add a reply

Find answers from the community

do you need an api key with GPT 4.0 to