Icon

2_​RAG_​ChatApp

<p><strong>RAG Chat App with source document Link</strong></p><ul><li><p>This Chat App connects to a vector store saved as .model file</p></li><li><p>if metadata contains a column "finalURL" with a link to a document, this is included in the chat bubble of the AI response</p></li></ul>

Build Vector Store

Connect to Azure OpenAI Chat Model

Finnovation Flows Retrieval Augmented Generation Data App

  • This workflow is a modification of the RAG Chat App created by KNIME. The Component has been upgraded with Python Views, which also work well on KNIME Hub and Community Hub.

  • The vector store should include a column in its meta data "finalURL" - this can store a link to the source document e.g. on SharePoint. If there is a document (or multiple) then links will be generated and included for each AI response (e.g. Source 1, Source 2...)

Provide your API key as Password

FF RAG ChatApp
Auth
OpenAI Authenticator
Pwd = API Key
Credentials Configuration
Model Reader
OpenAI Chat Model Connector

Nodes

Extensions

Links