Think inside your AI world.
Supabase through Unl
Supabase shows the schema and config. Whether a change is safe is a set of rules you decided — and the project doesn’t enforce your rules for you.
Supabase through Unl reads your live project — database, project management and config — against the safety rules you have ratified, so a proposed change comes back safe-or-not with the rule it breaks named.
What Supabase holds
Supabase’s MCP server groups its surface into feature areas (serving around 10M developers):
- Database schema and objects
- Project management
- Configuration
What the naked read gives you
A naked read returns the tables and settings. Accurate, and it won’t flag that a new table has row-level security off, because “every new table needs RLS” is your rule, not a Supabase default it applies to you.
The frame judges the data it is given; it does not verify the source’s accuracy.
What changes when Supabase is read measured
Say you’ve ratified two rules: every new table ships with row-level security on, and no destructive migration runs without a backup recorded.
“Is this migration safe?” returns, measured: not safe — the new orders table has RLS off, which breaks the rule you set so nothing goes live world-readable by accident. Supabase supplied the schema; your ratified rules supplied the check and the reason.
And back again
When you confirm a table is deliberately public and record that decision, the next safety read treats it as an allowed exception rather than a fresh breach.
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 check my Supabase changes against safety rules?
Yes, through Unl. You ratify the rules (say RLS on new tables, backups before destructive migrations, with your reasons), and the live project is measured against them — a change returns safe or not, with the rule it breaks named.
How do I connect Supabase to Claude?
Supabase ships an MCP server exposing database, project and config reads. Through Unl the same reads arrive measured against your ratified safety rules, so a change reads as a verdict rather than a schema dump.
Does Unl run migrations or change my database?
Unl reads through Supabase, and can write back on your explicit gesture — it never acts as a side effect of a read.
What if the backup was taken but recorded somewhere other than the database?
Then the rule reads as breached until the record is somewhere the read can see it. Supabase can show you the migration ran; whether a backup exists is a fact that lives wherever you put it. The read names the missing record rather than assuming a careful team took one, because a rule that assumes compliance is not a check.
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.