Think inside your AI world.
Metabase through Unl
Metabase runs the query and hands back the number. Whether that number is clean enough to put in front of the board is your rule - and 'the query ran' isn't the same as 'the data behind it is trustworthy.'
Metabase through Unl reads live query and dashboard results against the reporting-readiness criteria you have ratified, so 'can this number go in the board deck' comes back met-or-short with the condition it hasn't met.
What Metabase holds
What Metabase through Unl actually reads, direct from the Metabase MCP:
- Execution of constructed queries, with results, column metadata and row counts
- Paged query results across large tables via a continuation token
- Keyword and natural-language search across tables and metrics
- Metabase entities and lists read directly by their metabase:// URI
- Query construction against a specific table or metric, scoped to the connecting user's permissions
What the naked read gives you
Ask Metabase through Unl to run the quarterly-churn query and it comes back clean: the number, the row count, how long the query took, the columns it touched. That is the real, current answer the warehouse gave, scoped to exactly what the connecting person is allowed to see. What Metabase has no visibility into is whether that churn figure is built on a source table that finished its nightly sync, or one that's been stuck since Tuesday - the query still runs, and still returns a confident-looking number, whichever is true.
The frame judges the data it is given; it does not verify the source’s accuracy.
What changes when Metabase is read measured
Say you're an operations lead who owns the board reporting pack, and you've ratified that a metric is board-ready only if its source tables synced within the last twenty-four hours and the query has been reviewed by a second analyst in the last quarter - a fresh-looking number from a stale table doesn't qualify.
'Is the churn figure ready for the board deck?' returns, measured: not yet, because the source table last synced thirty-one hours ago, past the twenty-four-hour freshness your rule requires, even though the query itself is reviewed and current. Metabase supplied the result and the sync timestamp; your ratified rule supplied the freshness window.
And back again
Once the sync catches up and the number is re-pulled, log the result to Unl against your freshness rule, so the deck carries proof the twenty-four-hour bar was actually cleared - not just that someone eyeballed a dashboard and moved on.
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
Why did Metabase through Unl say this number isn't board-ready yet?
Because the table it's built from hasn't synced inside the freshness window you ratified, even though the query ran cleanly and returned a number. Once the sync completes, or the rule changes, the verdict updates with it.
How do I connect Metabase to Claude?
Your Metabase instance serves its own MCP endpoint at /api/metabase-mcp - add it with claude mcp add --transport http metabase https://your-metabase-url/api/metabase-mcp and authorise over OAuth. Unl reads through that same connection.
Does connecting Metabase let Unl edit my dashboards or questions?
Not through the read path Unl uses. Unl runs queries and reads results, scoped to the connecting user's own permissions - it doesn't rewrite dashboards or questions on your behalf.
What this is
Think inside your AI world — you stay in command
Unlimitless (Unl to friends) holds what you've settled, reads what your tools are showing, and catches what's changed out in the world — and hands your AI whatever bears on the work, the moment it's needed, without you asking. The right thing, in front of the model, unprompted, with you in command of the call. So you keep moving toward what you set out to build, on top of everything you've already decided.
It plugs into Claude, Claude Code, ChatGPT and Cursor as an MCP connector. Quick to connect, in a couple of steps.
Unlimitless is open now to invited Alpha. Apply for the Beta waitlist to come in ahead of the full launch:
Alpha is invite-only · free at launch.