Wiring AI agents to real enterprise data is harder than the demos suggest. The model part is easily solved: the frameworks are mature, the patterns are established, and the tooling is excellent. The data part is where things slow down. Every source has its own authentication model. APIs update versions without warning. Schema changes arrive silently. Rate limits compound. Each problem is solvable in isolation, but maintaining a constellation of custom connectors while shipping the thing you actually set out to build has a way of consuming projects before they launch.
The advent of AI-assisted coding could solve these problems, but it shouldn't. Even with LLMs at your disposal, your time and tokens are better spent on building the features and functions of your apps, agents, and platforms, not on plumbing. With a managed data layer that handles connectivity and infrastructure (source authentication, schema discovery and normalization, rate limits, pagination, etc.) for hundreds of sources, you can focus on building and shipping.
CData is easing onboarding onto the data layer that solves developers' plumbing problems and is already trusted by thousands of customers with three products: Connect AI Developer Edition (free), the Connect AI Python SDK (open source), and the CData CLI, all available now at cdata.com/developers.
Three releases, three on-ramps to one data foundation for AI
We are releasing three products rather than one because developers aren't working from a single surface. Gartner projects that 90% of enterprise software engineers will be using AI code assistants by 2028, up from less than 14% in early 2024 (Gartner). Some developers are wiring agents inside Claude Code or Cursor, where MCP-native tooling is the right interface. Some are building pipelines and internal tools in Python, where a DB-API cursor fits cleanly into existing workflow patterns. Others are using AI coding assistants that benefit from fast, terminal-accessible schema discovery without standing up a server at all.
No matter where they're working, developers (whether AI-assisted or AI-native) should focus on the code that powers their agents, apps, and platforms.
Connect AI Developer Edition: free enterprise data layer
Connect AI Developer Edition is the free tier of the CData Connect AI platform. With it, you get MCP-based access to every Connect AI source connector—Salesforce, Snowflake, NetSuite, Workday, Microsoft 365, and hundreds more—with support for up to five active connections, five users, and 500 tool calls per month. Sign up with any email. No credit card needed.
Developer Edition ships with the Query MCP server, Data Copilot, Toolkits, Connect Gateway, and the Management MCP server. The two built-in MCP servers offer different capabilities.
The Query MCP server has tools that let LLMs explore metadata, fetch live data, and act on the connected systems.
The Management MCP server (Endpoint: mcp.cloud.cdata.com/mcp/mgmt) exposes tools that let agents list available sources, create connections, and test them, enabling full setup from the IDE without opening a UI.
You can scaffold and configure a working data connection entirely from within your agent environment. Then use those connections to read and write data from live sources. Compatible clients include Claude Code, Codex, Cursor, GitHub Copilot, and LangChain. The tier is deliberately scoped for building and evaluating and when you're ready to deploy within the organization or to customers, the upgrade path to a paid plan is clear.
Sign up free at cloud.cdata.com/developer.
From sign-up to a working agent in under 10 minutes
There are three onboarding paths for Connect AI Developer Edition.
The first is Data Copilot: an AI assistant within Connect AI that can execute a connection setup and interact with enterprise data in about five minutes. Start free.
The second is the agent scaffold:
npx @cdatasoftware/create-agent
That command generates a working agent project wired to the Connect AI MCP server, with examples for the framework of your choice.
Lastly, the developer portal at cdata.com/developers covers guided paths for OpenAI, Anthropic, LangChain, and CrewAI; from quick starts through advanced integration patterns.
Connect AI Python SDK: DB-API access for programmatic workflows
The Connect AI Python SDK is a PEP 249 DB-API 2.0 client for the Connect AI Query API. It supports SQL access to all of your data through the standard cursor interface—execute, fetchall, fetchone—and works with pandas, SQLAlchemy, and any DB-API-compatible tool. Install it with:
pip install cdata-connect-ai
The SDK is MIT-licensed and available on PyPI and GitHub at CDataSoftware/cdata-connect-ai-python. It calls the Connect AI Query API, which requires a Connect AI account. The access pattern distinction is straightforward: Developer Edition is MCP-centric, suited for agent and IDE workflows; the Python SDK is built to power AI agents, ETL pipelines, BI notebooks, and internal tools where programmatic DB-API access is the primary need. And because it's based on an established API spec, your agents can build faster with it.
CData CLI: live schema access for AI coding assistants
The CData CLI (coming soon) is a terminal interface to CData's drivers and connectors. It lets AI coding assistants download drivers, discover schemas, inspect columns, and run SQL queries against live sources without standing up an MCP server. Install it with:
curl -fsSL https://downloads.cdata.com/cdatabuilds/builds/free/cdatacli/install-cdatacli-macos.sh | bash
Commands organize into four groups: drivers, connection, metadata, and query. All output is structured JSON — pipeable, scriptable, and agent-readable. The agent invokes the CLI, parses schema output, and generates source-specific queries without guessing at column names or types. With the CLI, developers can build with the assistance of AI and, by embedding CData drivers to power connectivity, ship without AI dependencies. Learn more, including the terminal commands to download and install the CLI, at cdata.com/solutions/cli.
The same governed foundation
Getting reliable access to enterprise data in an AI agent has historically meant building your own connectors, managing a patchwork of source-native MCP servers, or waiting for procurement. There was no free path in, no pip install that gave you Salesforce and Snowflake in the same cursor, no terminal command that surfaced live schema without standing up infrastructure.
That changes today. All three releases sit on the same connectivity and governance layer that is trusted by over 10,000 customers. A developer building in Developer Edition today uses the same source connectors, schema discovery, and authentication handling that a production enterprise deployment runs. The upgrade path when you're ready to deploy is a configuration change, not a replatform.
Which one is right for you?
| Connect AI Developer Edition | Connect AI Python SDK | CData CLI |
Best for | Building AI agents that need live access to enterprise data via MCP | AI agents, ETL pipelines, BI notebooks, and internal tools where programmatic DB-API access is the primary need | Developers using AI coding assistants to scaffold and query via CData drivers |
Interface | MCP — works with Claude Code, Codex, Cursor, LangChain, and any MCP-capable client | PEP 249 DB-API 2.0 — an established spec that works with pandas, SQLAlchemy, and standard Python tooling. Coding assistants generate correct code against it. | Terminal — structured JSON output, pipeable and agent-readable |
Cost | Free. 5 sources / 5 users / 500 tool calls per month. | MIT-licensed. Works with Developer Edition. | Free to install. Driver license required per source. |
Get started | cloud.cdata.com/developer | pip install cdata-connect-ai
| curl --fsSL .../install.sh | bash
|
Start building with governed enterprise data today
Connect AI Developer Edition gives you MCP-native access to Salesforce, Snowflake, NetSuite, Workday, Microsoft 365, and hundreds of other enterprise sources — free, no credit card, no connector to build. Sign up free at cloud.cdata.com/developer, or explore the full developer portal at cdata.com/developers to find quick starts for Claude Agent SDK, OpenAI Python SDK, LangGraph, and more.
Frequently asked questions
What's the difference between Connect AI Developer Edition and the Connect AI Python SDK?
Developer Edition is an MCP-centric free tier — it gives MCP-capable clients like Claude Code and Cursor access to enterprise data sources through a managed server. The Python SDK is a DB-API 2.0 programmatic interface to the Connect AI Query API for use in Python workflows. They serve different access patterns: MCP for agent and IDE workflows, DB-API for scripts and pipelines.
Does the Management MCP server come with every Connect AI plan?
The Management MCP server, which lets agents create and test connections without a UI, is included in Developer Edition. It exposes tools like list_available_sources, create_connection, and test_connection at mcp.cloud.cdata.com/mcp/mgmt.
What enterprise data sources can I connect to with Developer Edition?
Developer Edition gives MCP access to every Connect AI source connector, Salesforce, Snowflake, NetSuite, Workday, Microsoft 365, and hundreds more. The free tier supports up to five active connections. The full connector catalog is at cdata.com/ai/.
Is CData CLI free?
The CLI itself is free to install. Querying a specific data source requires a CData JDBC driver license for that source. Driver trials are available. Install with: curl --fsSL https://downloads.cdata.com/.../install.sh | bash
How do I upgrade from Developer Edition? What does a subscription look like?
When you reach the limits of Developer Edition (5 connections, 5 users, 500 queries), you'll be prompted to upgrade. You can upgrade from within the platform, on the Billing tab of the Settings page. Learn more about subscription plans at cdata.com/ai/pricing.
Explore the Developer Center today
Get the free Developer Edition of Connect AI and get your first answer in 5 minutes.
