Vapi is the Voice AI platform for developers.

Vapi AI

Hi can you please tell me how rag is used in vapi.what is the most efficinet way on collecting data in kb.How can i implememt rag

<@1296693150844190812>
It seems like there might be a bit of confusion in your query. "RAG" is not a term directly associated with VAPI (Voice API). If you're referring to a specific concept or feature related to VAPI that is abbreviated as "RAG," could you please provide more details or clarify what you mean?

Regarding the efficient way of collecting data in knowledge bases (KB), VAPI itself is focused on voice-related functionalities and may not directly handle knowledge base management. However, if you're looking to integrate VAPI with a knowledge base for voice applications, you would typically:

1. Define the structure of your knowledge base to ensure it's optimized for voice queries.
2. Use VAPI to interpret user voice inputs and convert them to text.
3. Query your knowledge base with the processed text to retrieve relevant information.
4. Use VAPI to convert the retrieved text data back into voice responses for the user.

If you need specific guidance on implementing a feature or if "RAG" refers to something specific within the context of VAPI, please provide additional details so I can assist you more accurately. If this is a feature or concept not covered in the VAPI documentation, you might need to reach out to support for more information.

You can continue this conversation by mentioning me in the message.

<@1296693150844190812> checkout this document on using knowledge base https://docs.vapi.ai/knowledgebase and we have our RAG template which we use to server the content during the call.

For most efficiency use KB with markdown formatting and text type only.

Do let me know how can i help you further.

<@865531204849500190> I want to ask and confirm if vapi uses RAG?

<@735161176454987807> VAPI uses RAG template to make use of content returned by the Knowledge Base.

<@865531204849500190> Thank you very much for your answer. Could I know what RAG template do you use?

we use ragTemplate function which formats context and queries into our template and then added as user message into the conversation that includes parsed KB content.

```
template
{content}

instructions

Query: {userQuery}
```

ragTemplate function formats a query and relevant context into a structured template to guide LLM in generating a response based solely on the provided information, while avoiding the use of prior knowledge.

Thank you so much for the support and answering <@865531204849500190>