Chat

Multi-model AI chat with SQL, RAG, and data analysis capabilities.

Overview

Chat Studio is a conversational AI interface backed by the Calliope data agent. It supports natural-language queries against connected databases, RAG over your documents, and general-purpose AI conversation — all in one unified chat window. You can switch between AI providers and models on the fly.

Key Features

• Multi-model support — AWS Bedrock (Claude, Titan), OpenAI, and more, switchable per session
• SQL agent — Ask questions in plain English, get SQL queries and results back
• RAG — Upload documents and ask questions against them with vector search
• Demo databases — Pre-connected sample databases to explore immediately
• Streaming responses — Answers stream in real time
• Session history — Conversations persist across sessions

Getting Started

1. Launch Chat from the Hub
2. Select your AI model from the model picker
3. Type your question or data request
4. For SQL queries, make sure a datasource is selected in the sidebar

SQL Queries

Chat can query your connected databases directly. Just ask in plain English:

Show me total revenue by region for 2024

The agent:

1. Inspects your schema automatically
2. Generates the SQL
3. Executes it against your database
4. Returns results as a formatted table
5. Offers a visualization if the data is chart-friendly

Demo Databases

Five sample databases are pre-connected for immediate exploration:

RAG (Document Q&A)

Upload documents and ask questions against them:

1. Click the Documents section in the sidebar
2. Upload PDFs, Markdown, or text files
3. The system indexes them with vector embeddings
4. Ask natural language questions — relevant context is retrieved and included automatically

Switching Models

Use the model picker to switch AI providers:

• AWS Bedrock — Claude Haiku, Sonnet, Titan (default)
• OpenAI — GPT-4o, GPT-4 Turbo
• Additional providers configured by your organization

When to Use Chat vs Other Tools