Explore and analyze your data with OpenAI models

Connect AI powers agents that can securely read, write, and act on live data across your warehouses, lakehouses, and hundreds of enterprise systems.

Get Started Free

Trending sources for ChatGPT:

Popular ChatGPT integrations

CData is OpenAI certified for agent connectivity

ChatGPT App Directory

Use live data from your warehouses and databases directly in ChatGPT—with virtual datasets structured around how your business asks questions, not how your storage was designed.

ChatGPT Agents

Build autonomous agents with the OpenAI Agents SDK connected to live enterprise data. Run multi-step analytical and operational tasks with governed, real-time data context.

OpenAI API

Access MCP-enabled data feeds directly from the OpenAI API to build high-context custom integrations, pipelines, and data-aware applications.

Give OpenAI a curated, controlled view of your data—not raw database access

Step 1:

Connect your data sources CData applies a standardized relational layer to every connected source—no data movement, no exports, no code required.

Step 2:

Define virtual datasets and build agent toolkits Administrators define virtual datasets across sources, encode relationships, and apply business context. These are bundled into agent toolkits and exposed to ChatGPT as dedicated MCP Servers.

Step 3:

Configure governance Users authenticate directly to the source system, with additional downscoped permissions layered on top. Access policy is defined in CData and enforced at run time—not delegated to the user or the model.

Step 4:

ChatGPT queries your data through CData CData federates the query, joins across sources, applies row- and column-level filters, and returns only what the user is authorized to see. Processing happens before results reach the model—minimizing token usage. Every interaction is logged: who asked, what was returned, when.

Looking for conversation ideas in ChatGPT?

Prompt Library

Explore our AI prompt library for ideas on next-level analysis you can do with AI + connected data.

Explore AI prompts

SDKs

Build production ready AI agents yourself using Python, the OpenAI Agent SDK, and CData Connect AI

Build production-ready AI agents with OpenAI

Get started in minutes using sample data in our step-by-step ChatGPT setup guide

Try it yourself

Explore the difference Connect AI makes for OpenAI

MCP connectivity that goes beyond the OpenAI ecosystem

Managed MCP with Connect AI

Individual user MCP connections

One governed MCP connection to hundreds of enterprise sources—cloud data warehouses, relational databases, data lakehouses, CRMs, ERPs, and more—under a single access-controlled layer.

Each data source requires a separate MCP server with its own tools, authentication and policies.

A standardized relational layer that returns data in a consistent semantic format ChatGPT can interpret reliably—pushing computation to the query engine, not the LLM, so token usage stays lean.

ChatGPT must filter, join, and group datasets logically through code, consuming up to 10X more tokens and ballooning costs in some cases

Purpose-built and custom MCP tools—universal tools for common queries across any connected source, plus the ability to define custom tools and toolkits scoped to your specific datasets and workflows.

Varied based on each individual API, adding to MCP context bloat and overwhelming ChatGPT thinking windows

Centralized governance control plane—user identities mapped to source system credentials, permissions downscoped at the dataset level, and a full audit trail across every agent and user query.

Connections are configured ad hoc by users, with limited visibility or centralized access governance.

Learn more about connected data for ChatGPT

Read our full announcement to see how Connect AI handles connectivity across sources, context engineering for enterprise data, and governance that IT can actually enforce.

Read the blog

Start querying data from ChatGPT within minutes

Start a 14-day free trial and see the difference connected data makes in ChatGPT yourself

Get a Free Trial

CData is the data layer that makes AI work in production—live connectivity and replication across 350+ sources, semantic context, and built-in governance. Powering AI for Databricks, Microsoft, Google, Palantir, and 10,000+ customers worldwide.