Endpoint Selection
This section outlines the steps and best practices for integrating and utilizing RAG-Buddy within your RAG pipeline. The first step is to choose the right endpoint.
API Endpoint types
This section will help you choose the right endpoint. RAG-Buddy currently offers two type of endpoints:
- Proxy Endpoints, start with
https://api.ragbuddy.ai/proxy - REST API Endpoints, start with
https://api.ragbuddy.ai
API Endpoints
As we currently support two use cases for integrating RAG-Buddy into your LLM pipeline, we offer two endpoints per endpoint type:
- Proxy endpoint for RAG+C:
https://api.ragbuddy.ai/proxy/ragc - REST API endpoint for RAG+C:
https://api.ragbuddy.ai/ragc - Proxy endpoint for Text-classification:
https://api.ragbuddy.ai/proxy/tc - REST API endpoint for Text-classification:
https://api.ragbuddy.ai/tc
Understanding the Differences
Proxy Endpoint
The Proxy Endpoint acts as a direct substitute for your existing LLM URL. It’s designed to intercept calls to the LLM and augment them with RAG-Buddy capabilities.
REST API Endpoint
The REST API Endpoint is a more versatile and powerful interface. It allows explicit data handling and offers more control over how you interact with RAG-Buddy. The response from the endpoint is a JSON object with a predictable data structure.
Pros and Cons
Proxy Endpoint
- Pros: Easy to implement, minimal changes required in existing setup.
- Cons: Less flexibility, reliant on the structure of the original LLM calls. More difficult to adhere to the required message template format.
REST API Endpoint
- Pros: Greater control, allows for more complex interactions and is less sensitive to LLM messages formatting.
- Cons: Requires more setup and understanding of the API.
Recommended for Production
For production environments, we recommend using the REST API Endpoint. It offers a robust set of features and flexibility, allowing you to fully leverage the capabilities of RAG-Buddy in diverse scenarios.