typesense

Hi eveyone

I am trying to build a chatbot that is connected to typesense RAG. I have a products collection but when I am trying to do earch with additional parameter conversation=true I am getting this message
Conversation mode requires an auto-embedding field to be specified as a Query By field.
I understand I have to create an embedding column in my collection schema but it is asking for model
```"model_config": {
          "model_name": "ts/e5-small"
        }```
how can I decide what model to use? what are my options here?

Hi <@U099Z53U552>,

The RAG functionality performs a vector search on your documents. Therefore, you'll need to set up your document embeddings first! You can read more about it here:

<https://typesense.org/docs/29.0/api/vector-search.html#index-embeddings>

The model_config is where you decide wheter you want to use a builtin model or a external provider like OpenAI.

The differences will be the model potency and, of course, the capacity for generating these embeddings.