Think inside your AI world.
Snowflake through Unl
Snowflake’s Cortex answers questions in plain language. Whether the answer uses your definition is still your call — the semantic model has its own.
Snowflake through Unl reads your live data via Cortex Analyst and Cortex Search, then holds the semantic answer to the metric meaning you ratified — so a Cortex answer arrives on your definition, and where the semantic model’s own logic diverges, that difference is named.
What Snowflake holds
Snowflake’s managed MCP server exposes semantic querying:
- Cortex Analyst (natural-language querying)
- Cortex Search (semantic search over data)
What the naked read gives you
A naked read lets Cortex answer from its semantic model. Powerful, and it answers on the model’s definitions, not necessarily the one you ratified — a difference that hides inside a confident answer.
The frame judges the data it is given; it does not verify the source’s accuracy.
What changes when Snowflake is read measured
Say you’ve ratified how revenue is recognised for the metric: net of refunds, booked at delivery, in the reporting currency — the definition finance signed off.
“What was revenue last quarter?” returns, measured: on your ratified basis — net of refunds, booked at delivery — the figure is £1.18m. Cortex, left to its own semantic model, books gross at invoice and answers £1.31m; that £130k is not an error but a different meaning, and the meaning that governs here is the one finance signed off.
And back again
When you align the semantic model to the ratified basis, that alignment is recorded in Unl, and from then on Cortex answers are read against the meaning finance holds, not the model’s default.
The answer comes back measured against what you already decided, and why.
The lane is live and open to this tool today: 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.
Questions people ask
Can AI hold Snowflake's answers to our metric definition?
Yes, through Unl. You ratify the definition, and the Cortex answer is measured against it — the number holds your meaning, and any divergence from the semantic model’s own logic is flagged.
How do I connect Snowflake to Claude?
Snowflake offers a managed MCP server with Cortex Analyst and Cortex Search. Through Unl the same semantic answers arrive measured against your ratified definition, so an answer reads on your terms.
Does Unl change my Snowflake data?
Unl reads through Snowflake, and can write back on your explicit gesture — it never acts as a side effect of a read.
What if a refund lands after the period has already been reported?
Then the figure that was reported and the figure the rule now returns are different, and the read shows both rather than silently restating. Your definition is net of refunds, so a later refund does move the number; whether the reported period is reopened is a finance decision, not a query result. The read gives you the movement and the date it happened.
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.