Think inside your AI world.

Axiom through Unl

Axiom's MCP server will run the query and show the error rate spiking. Whether that spike is worth waking someone up for is a threshold your on-call rotation has already argued about once - a spike that crosses a line isn't the same as one that holds.

Axiom through Unl reads live error-rate query results against the incident-escalation threshold you have ratified, so 'is this actually an incident?' comes back clear-or-flag with the condition it hasn't met.

What Axiom holds

Query Axiom's MCP server about a dataset and it returns:

  • live APL query results against any dataset
  • dataset list and schema for a given dataset
  • monitor status checks across all monitors
  • alert history for a specific monitor
  • saved APL queries available for reuse

What the naked read gives you

Axiom's MCP server runs the APL query, lists the datasets, checks every monitor's current status, and pulls the alert history behind any of them. It reports exactly what the telemetry says happened, down to the minute, without judging whether what happened rises to the level of waking someone up.

The frame judges the data it is given; it does not verify the source’s accuracy.

What changes when Axiom is read measured

An on-call SRE at a payments company ratified an escalation rule in Unl: an error-rate spike only pages the on-call rotation once it exceeds two percent sustained for ten minutes, because false pages at 3am for four-minute blips had burned out two engineers before the threshold existed.

'Is this actually an incident?' returns, measured: not yet - the APL query shows the checkout-service error rate crossed two percent, but the spike lasted four minutes before recovering, short of the ten-minute sustain the rule requires. Axiom supplied the query result; your ratified rule supplied the escalation threshold.

And back again

If the error rate spikes again and holds past ten minutes, Axiom through Unl catches the extended duration on its next look; the ratified escalation rule pages on-call then, and the verdict flips on this same page.

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

What would turn this into a real incident?

The elevated error rate holding for the full ten minutes rather than dropping back down. The same ratified escalation rule reads the sustained duration and pages on-call once it does.

How do I connect Axiom to Claude?

Authorise Axiom's remote MCP server via OAuth in Claude's MCP settings - credentials stay isolated and revocable - or run it locally against your Axiom API token. Unl reads through that connection.

Does Unl write to my datasets?

No. Unl calls queryApl, checkMonitors, and related read tools to see current state; it creates no dashboards or monitors on your behalf. Any escalation, your paging system handles it.

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.