Think inside your AI world.
How do I connect Databricks to ChatGPT?
Databricks' managed servers — Unity Catalog, AI Search, Genie spaces, custom functions, AI Gateway tools — serve ChatGPT the governed lakehouse.
Direct answer: connect Databricks' managed MCP servers to ChatGPT. In ChatGPT, add it from the connector settings (developer mode); the tool's own docs carry the current path. Authorise the workspace and ChatGPT reaches Unity Catalog assets, AI Search, Genie's natural-language analytics spaces, custom functions, and the AI Gateway's SaaS tools — inside the governance the workspace already enforces.
Governed answers, conversational route
Which assets serve the customer domain, what Genie returns for the revenue question, what the function computes — answered through managed servers that inherit workspace permissions rather than tunnel under them.
Notebooks are drafts until reviewed
Worked case: your team ratified a notebook-to-prod rule — no scheduled job runs from an unreviewed notebook, after a prototype with a hardcoded date quietly wrote a month of wrong aggregates. Databricks schedules whatever you point it at; that review stands between prototype and schedule is engineering law you wrote over the scheduler.
Through Unl the pipeline read arrives with the law held: one scheduled job traces to a notebook with no review record — against your rule, the wrong-aggregates month attached. Drafts stay drafts until your process says otherwise.
Databricks to ChatGPT reads the governed lakehouse; through Unl the lakehouse answers to the promotion laws you ratified — the unreviewed notebook gets stopped at the schedule, not discovered in the aggregates.
Reads through Unl arrive with measured context — in the presence of the decisions you’ve already settled. The reach lane is live: one box, paste anything. If it speaks MCP, Unl can reach it. Readings arrive unprompted, the data beside the criterion; Unl is a courier, not a warehouse, and keeps only your keys and the frame.
Read further
Questions people ask
What do the managed servers expose to ChatGPT?
Unity Catalog, AI Search, Genie spaces, custom functions and AI Gateway SaaS tools — managed, permission-inheriting, identical across MCP clients.
How does this compose with Unity Catalog?
Catalog governs data access; Unl governs your operational judgement above it — what may be scheduled, promoted or retrained, and why. Two layers, no overlap.
What promotion rules ratify well?
Review-before-schedule, quality-gates-before-retrain, named ownership for production jobs. Each has an incident; each earns its keep at read time.
What if the review becomes a rubber stamp?
Then you have the ceremony without the check, and the record will not save you because it shows approvals either way. What makes a review real is that it can fail. If nothing has ever been sent back, that is worth asking about before the aggregates make you ask.
What this is
Think inside your AI world — you stay in command
Save the thoughts, decisions and targets worth keeping, each with its reasoning, carried into every AI session the moment they matter. A new unit of exchange between you and your AI: the Settled Why with standing that travels. Unprompted.
Your whole AI world. What you decided at the epicentre. It plugs into Claude, Claude Code, ChatGPT and Cursor as an MCP connector — quick to connect, in a couple of steps.
MCP native·Human settled·Model agnostic·Your data
Measured Context
Connect Unl to bring the right information into the moment.
Your sources, read against the criteria you set.
The full product, open. Free at launch.