Think inside your AI world.
Statsig through Unl
Statsig shows the experiment lift and the gate status. Whether that lift is real enough to roll out to everyone is your rule - and a green metric on day two isn't the same as a green metric that held for two weeks.
Statsig through Unl reads experiment, gate and metric data against the significance and duration criteria you have ratified, so 'is this experiment ready to ship' comes back clear-or-hold with the condition it hasn't met.
What Statsig holds
What Statsig through Unl actually reads, direct from the Statsig MCP:
- Feature gate configuration and current rollout status
- Experiment (A/B test) results and variant performance
- Dynamic configs and layers assigned to a project
- Metrics and the metric sources feeding them
- Segments and audit log entries for a given change
What the naked read gives you
Ask Statsig through Unl to pull an experiment and it hands back the real result: the lift on the primary metric, the confidence interval, which variant is winning, how the gate is rolled out today. None of that is dressed up - it is the actual number Statsig computed from the actual traffic that hit the test. What Statsig has no way of knowing is whether a 4% lift with a confidence interval that only just clears zero, measured over four days, is the kind of result this particular team ships on, or the kind they wait out for another week. The dashboard shows green. Whether green is enough is not a Statsig question.
The frame judges the data it is given; it does not verify the source’s accuracy.
What changes when Statsig is read measured
Say you're a platform engineer who owns the checkout experiment programme, and you've ratified that a result only ships to 100% once the primary metric has run for at least fourteen days and the confidence interval sits entirely above zero - a promising early trend on its own does not count.
'Is the new checkout flow ready to ship?' returns, measured: not yet, because the experiment has run nine days against the fourteen-day minimum, even though the lift itself already clears the interval your rule requires. Statsig supplied the lift, the interval and the day count; your ratified rule supplied the fourteen-day floor.
And back again
Once the experiment crosses day fourteen, the next Statsig through Unl read writes the outcome back to Unl against the same rule - so the rollout decision and the reason it finally cleared are both on record, not just the fact that someone eventually clicked ship.
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 Statsig through Unl say this experiment isn't ready to ship?
Because it hasn't run long enough against the rule you ratified, even though the lift itself already looks good. Statsig reported nine days of data against a fourteen-day minimum; once the clock catches up, or the rule changes, the verdict changes with it.
How do I connect Statsig to Claude?
Enable the Statsig MCP server for your project and connect over OAuth to the hosted endpoint at api.statsig.com/v1/mcp. Unl reads through that same connection once it's authorised - no separate integration needed.
Does connecting Statsig let Unl change my gates or experiments?
No. A read-only connection only exposes gates, experiments, configs, layers, metrics and audit logs for querying - it doesn't create, edit or ship anything. Unl reads one way, into the verdict.
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.