Skip to main content
Version: v1.7.3-14 🚧

Chats

Overview​

This section provides different actions you can take around chats in Enterprise h2oGPTe. For example, create a chat session with or without a Collection or an imported document(s).

Types of chats​

Collection-backed chats​

A collection-backed chat connects your conversation to a document collection. The LLM answers questions using your uploaded documents as context (RAG: Retrieval-Augmented Generation). Use this when you want answers grounded in specific files, knowledge bases, or data sources.

Chats without a collection​

A chat without a collection lets you converse directly with an LLM without selecting any documents. The model responds without document context. When agents are enabled, the model can also use tools such as web search, code execution, and file generation. Use this for general Q&A, brainstorming, summarization of pasted text, or coding assistance.

note

Creating a chat without a collection requires the Create a chat without a collection permission. See Roles and Permissions for details.

Showcase​

Administrators and users with the Manage Showcase permission can highlight selected shared chats on a curated Showcase page. The Showcase displays featured conversations as a tile grid with custom titles, descriptions, and thumbnails.

AI-powered chatbots​

Enterprise h2oGPTe allows you to build AI-powered chatbots. To learn more, see Tutorial 2: Build an AI-powered chatbot (model) to enhance a website's search capabilities.

Tutorial 2: This tutorial with Enterprise h2oGPTe and the h2oGPTe Python Client Library builds an AI-powered chatbot to replace the function of a website's search bar, which, in turn, builds something better to enable users to obtain better answers to their questions about the website. In this tutorial, we will create an AI-powered chatbot to enhance the search capabilities of the H2O Model Validation documentation website.

Feedback