Goals
- objectives
- KRs
- bets
What you're trying. Why it matters. What you'd ship to find out.
Signals you trust. Guardrails you set. Decisions you can defend.

What you see, day one.
Pull signals from the tools you already trust — GitHub, Datadog, Sentry, Stripe — alongside the ones Merak generates itself. Watch them next to the KRs they roll up to.

Live signals · sample
P95 latency · /checkout
via Datadog
Bug rate · last 7 days
via Sentry
Cost / active user
via Stripe
Activation · 7-day
via Merak
PR review latency
via GitHub
Same numbers humans and agents read from. KRs live next to the signals they roll up to.
What we replace.
Standups, OKR docs, status packets, RCA theater. None of it scaled to AI-native operations. None of it should have.
The substrate.
Goals on top, work in the middle, truth underneath. The same primitives humans use, exposed to the agents you let in.
What you're trying. Why it matters. What you'd ship to find out.
How it gets done — humans, agents, or both.
What's actually happening. Pulled from your tools. Generated by Merak. The same numbers humans and agents read from.
What beta looks like today.
No mockups. No "coming soon" gradients. Same UI you see five minutes after sign-in.
← signals attached to live KRs

Connect what you already use
Pricing.
Free during open beta. No annual contracts, no per-feature gates, no enterprise tax.
$0
180 agent minutes / month · ≈ 6 experiments
$29
1,500 agent minutes / month · ≈ 50 experiments
Need limits, FAQ, cancellation, tax info? See full pricing →
What we ship today. Honest controls, not compliance theater.
TLS everywhere. Provider tokens stored as encrypted credentials.
Your workspace content is not used to train any model.
Owner-only delete and core data export are in progress for beta.
Read the full posture on Trust →
Open Beta
Free during early access. Pricing locks at $29/seat at GA. We respond within a day.