Think inside your AI world.
Confluence through Unl
Confluence shows the page and when it last changed. Whether that page is the approved version of the spec is your call - a fresh edit isn't the same as a signed-off one.
Confluence through Unl reads pages, spaces and search results against the approval criterion you have ratified, so 'is this the approved spec' comes back green-or-flag with the condition it hasn't met.
What Confluence holds
Confluence's official MCP server exposes:
- page search and retrieval across the spaces you can access
- space listing - 'what spaces do I have access to'
- page content summarisation on request, eg 'summarise the Q2 planning page'
- read and search permission groups alongside write, via OAuth 2.1 or API token
What the naked read gives you
Point Confluence's own MCP at the payments spec and it returns the page, the space it lives in, and whoever last touched it. That's an accurate read - the content and its edit history are right there. What it can't tell you is whether the page in front of you is the one engineering signed off, or a draft somebody reopened last Tuesday. Confluence logs every edit precisely; deciding which edit counts as the approved one sits outside what an edit history can tell you.
The frame judges the data it is given; it does not verify the source’s accuracy.
What changes when Confluence is read measured
Your engineering manager ratified the rule in Unl: a spec only counts as approved once it carries the 'Reviewed' label and is linked from the current release page - because an edited page with neither is still somebody's draft.
'Is the payments spec approved' returns, measured: not met - the page was edited eight hours ago but carries no 'Reviewed' label and isn't linked from the release page yet. Confluence supplied the page content and edit history; your ratified rule supplied the approval criterion.
And back again
Once that judgement is ratified, Unl carries a fixed definition of 'approved' for this spec, so the next engineer who opens it inherits the bar you set rather than inferring readiness from an edit timestamp.
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 Confluence tell me if a spec is actually approved?
Confluence's MCP returns the page, its space and its edit history; it doesn't decide what counts as approval. That's the criterion you ratified in Unl, which turns the page read into a verdict.
How do I connect Confluence to Claude?
Atlassian runs an official remote MCP server for Confluence, hosted on Cloudflare infrastructure - you authenticate with OAuth 2.1 or an API token and Claude connects directly, with no local install required.
Does connecting Confluence let Unl edit my pages?
No - Unl reads the pages, spaces and search results Confluence's MCP exposes. It doesn't publish or change page content on your behalf.
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.