Skip to main content

Connect DataHub Docs to AI Tools

Query DataHub docs directly from your AI assistant. We publish a machine-readable index of the documentation so AI coding tools can return accurate, current answers about DataHub — without you switching tabs.

What We Publish​

DataHub maintains an llms.txt file: a machine-readable index of the documentation, written for AI tools.

https://docs.datahub.com/llms.txt

Quick Start by Tool​

Cursor​

Add DataHub docs as a custom source:

  1. Open Settings → Features → Docs
  2. Click + Add new doc
  3. Enter https://docs.datahub.com as the URL

Cursor will index the site. Reference it in chat with @DataHub.

Claude Code​

Reference the index in your prompts:

claude "Using https://docs.datahub.com/llms.txt as reference, how do I set up DataHub ingestion from Snowflake?"

Or add the URL to your project's CLAUDE.md so Claude Code uses it on every turn.

Claude (Web & Desktop)​

Paste the llms.txt URL into the chat:

Use https://docs.datahub.com/llms.txt as reference. How do I write a custom ingestion source in DataHub?

Claude will fetch the index and the relevant linked pages.

ChatGPT​

With browsing enabled, paste the llms.txt URL into your chat. ChatGPT will use it as a navigation aid for the rest of the conversation.

GitHub Copilot (VS Code)​

In VS Code, reference DataHub docs in Copilot Chat using #fetch:

#fetch https://docs.datahub.com/llms.txt explain DataHub's metadata model

Beyond Docs: AI Access to Your Data Context​

Connecting AI to docs is one layer. DataHub also provides AI access to your metadata:

  • MCP Server — Plug Claude, Cursor, or any MCP-compatible client directly into your DataHub instance. Query lineage, find PII, search assets in natural language.
  • Agent Context Kit — Pre-built integrations for LangChain, Cursor, Claude, Gemini CLI, Vertex AI, Snowflake Cortex, Databricks Genie, and Microsoft Copilot Studio.
  • Ask DataHub (Cloud) — Natural-language search across your metadata.
  • Analytics Agent — Open-source agent (Apache 2.0, bring your own LLM) that turns plain-English data questions into SQL, results, and charts — grounded in your DataHub catalog.

Feedback​

This is an early step toward making DataHub docs first-class for AI workflows. If your AI tool isn't covered above or you have ideas for what to add next, let us know: