CData Connect AI MCP Server

Managed remote MCP server from CData that exposes 350+ enterprise data sources (Salesforce, Snowflake, NetSuite, etc.) through a unified SQL interface.

Data & Enrichment by CData Software OAuth2 or Basic Auth (Personal Access Token) active

Connect (Remote Server) Docs

Overview

CData Connect AI is a fully managed, provider-hosted MCP server that gives AI agents real-time, read/write access to enterprise data systems through CData's library of 350+ connectors. Instead of running a separate MCP server per source, agents connect to a single remote endpoint and CData normalizes every backend (databases, SaaS apps, file stores, APIs) into a standardized SQL layer that LLMs can introspect and query.

The server exposes a small, universal toolset for catalog discovery, schema and table introspection, SQL query execution, and stored procedure invocation. Stored procedures cover non-SQL operations like uploading attachments, sending messages, or closing tickets in source systems. Because CData enforces the underlying connection credentials and permissions, data stays in place and access respects the original source's authorization model.

Notable for being one of the few official, vendor-maintained managed MCP platforms covering a very broad set of enterprise sources, with support for Claude, Microsoft Copilot Studio, n8n, and any other MCP-compatible client over Streamable HTTP transport.

Tools

Tool	Description
`getCatalogs`	Lists the available connections (catalogs) configured in Connect AI.
`getSchemas`	Retrieves the schemas available within a specific catalog.
`getTables`	Lists tables for a given catalog and schema.
`getColumns`	Returns column definitions and types for a specified table.
`queryData`	Executes a SQL query against one or more connected data sources.
`getProcedures`	Retrieves the list of stored procedures available in a catalog and schema.
`getProcedureParameters`	Returns parameter metadata for a specific stored procedure.
`executeProcedure`	Runs a stored procedure with the supplied parameters (e.g. upload attachment, close ticket).

Setup Guide

CData Connect AI is a remote, managed MCP server. There is nothing to install locally. You add the endpoint URL to any MCP-compatible client and authenticate with either OAuth (handled by the client) or a Personal Access Token issued from the Connect AI Settings page.

Prerequisites

A CData Connect AI account (sign up at https://www.cdata.com/ai/)
At least one configured data source connection in Connect AI
A Connect AI User ID and Personal Access Token (PAT) generated from Settings, if using Basic Auth

Remote MCP URL

https://mcp.cloud.cdata.com/mcp

Claude Desktop / Claude.ai (Custom Connector) Add a Custom Connector and paste the Remote MCP Server URL above. Choose OAuth or Basic Auth and supply your User ID and PAT when prompted.

Generic MCP client config (Streamable HTTP)

{
  "mcpServers": {
    "cdata-connect-ai": {
      "url": "https://mcp.cloud.cdata.com/mcp",
      "transport": "http",
      "headers": {
        "Authorization": "Basic <base64(UserID:PAT)>"
      }
    }
  }
}

Other supported clients

Microsoft Copilot Studio: paste the URL into the Remote MCP Server URL field
n8n: configure the MCP node with the same URL and credentials

Full reference: https://docs.cloud.cdata.com/MCP.html

Use Cases

Let an agent query Salesforce, NetSuite, and Snowflake from a single MCP endpoint without standing up per-source servers.
Allow a Claude agent to discover schemas and run ad-hoc SQL across enterprise apps using natural language ("show me top accounts by ARR from Salesforce joined with usage from Snowflake").
Trigger source-system actions through stored procedures, e.g. close a Zendesk ticket, send a Slack message, or upload a file to SharePoint.
Power Microsoft Copilot Studio or n8n workflows with governed, credentialed access to operational data sources.
Build internal AI assistants that read and write to back-office systems (ERP, CRM, HRIS) while keeping data in place and respecting existing permissions.

Example Prompts

"List all available catalogs, then show me the tables in the Salesforce catalog."
"Query Snowflake for the top 20 customers by revenue in Q1 and enrich them with the Industry field from NetSuite."
"Find all open opportunities in HubSpot owned by Jane and update their stage to 'Closed Lost'."
"Run the closeTicket stored procedure on the Zendesk connection for ticket ID 48211."
"What columns are available on the Orders table in the Shopify catalog?"

Pros

Official, fully managed remote server: no local install, no per-source MCP setup.
Broad coverage of 350+ enterprise data sources through CData's mature connector library.
Read and write support, including non-SQL actions via stored procedures.
Standardized SQL interface across heterogeneous backends, making LLM queries consistent.

Limitations

Commercial product: requires a CData Connect AI subscription, not free or open source.
All access flows through CData's hosted infrastructure, which may not fit strict data-residency requirements.
SQL-centric interface means non-SQL semantics of some SaaS APIs must be expressed through stored procedures.

Alternatives

CData Connect Cloud MCP Server (self-hosted CData option)
MCP Toolbox for Databases (Google's open-source DB-focused MCP)
Composio or Pipedream MCP gateways (alternative managed multi-source MCP platforms)