Think inside your AI world.

LaunchDarkly through Unl

LaunchDarkly shows the flag's status in every environment. Whether it's safe to delete is your rule - and 'nobody's touched it in a sprint' isn't the same as 'nobody depends on it.'

LaunchDarkly through Unl reads flag status, targeting and evaluation history against the retirement criteria you have ratified, so 'is this flag safe to remove' comes back sound-or-flag with the condition it hasn't met.

What LaunchDarkly holds

What LaunchDarkly through Unl actually reads, straight from the LaunchDarkly MCP:

  • Feature flag status and targeting rules across every environment
  • Full flag listings and individual flag configuration
  • AI Config targeting, variations and listings
  • Audit log entries tied to a flag or project
  • Observability data - logs, traces and error groups linked to a flag

What the naked read gives you

Point LaunchDarkly through Unl at a flag and it lays out the truth of it: on in production, off in staging, a targeting rule that hasn't shifted in eleven weeks, an audit trail showing the last edit was a typo fix. That is a complete and honest picture of the flag's current state. What it leaves for a human to decide is whether eleven weeks of quiet targeting means the flag is dead code waiting to be deleted, or a kill switch nobody has needed to touch yet - which is exactly the flag you do not want someone deleting on a Friday.

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

What changes when LaunchDarkly is read measured

Say you're an engineering lead running quarterly flag cleanup, and you have ratified that a flag counts as safe to retire only when it has shown no targeting changes for eight weeks, evaluates to the same variation for every user, and isn't tagged as a kill switch in its description - the quiet ones with a kill-switch tag stay, regardless of how quiet.

'Can we delete the legacy-checkout flag?' returns, measured: not yet, because although it has been static for eleven weeks and evaluates to one variation, its description carries the kill-switch tag your rule excludes. LaunchDarkly supplied the targeting history and the evaluation pattern; your ratified rule supplied the kill-switch exception.

And back again

Retire the flag once the tag is gone and next quarter's check confirms the pattern held; log that outcome to Unl against your rule, so the record shows the retirement criteria were actually applied to this flag - not just that someone remembered to look.

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 LaunchDarkly through Unl say this flag isn't safe to retire yet?

Because it's tagged as a kill switch, and your ratified rule excludes kill switches from retirement regardless of how quiet their targeting has been. Remove the tag, or change the rule, and the verdict moves with it.

How do I connect LaunchDarkly to Claude?

Use LaunchDarkly's hosted MCP server, reachable at mcp.launchdarkly.com, and authorise it for your account from the Getting Started page. Unl reads flags, configs and observability data through that same connection.

Does connecting LaunchDarkly let Unl delete or edit my flags?

No. Unl reads flag status, targeting and audit history to build the verdict - it doesn't create, edit, archive or delete a flag 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.