Insure what breaks. Not who answers.
Every other pager bills you by the head, so the new engineer joining the rotation is a line item. We price what you're protecting instead — the users, the rotations and the escalation policies are free, however many of each you have.
- 1 service, round-the-clock
- Unlimited users
- Slack, email, push, webhooks
- SMS & voice paging
- SSO, SCIM, audit log
- Unlimited users, always free
- Unlimited rotations & integrations
- Day or round-the-clock cover
- SMS & voice, pooled then metered
- SSO, SCIM, audit log
What counts as a protected service
This is the whole billing surface, so it lives here in plain sight rather than in the terms. You should be able to count your own bill before you sign up.
A protected service is an enabled service with an escalation policy attached.
If it can wake someone up, it's insured — and you're paying for the guarantee that it will. Everything else about your org is free to grow: the same list of services costs the same whether four engineers cover them or forty, on one rotation or nine.
Not everything deserves a 3am phone call
So you don't pay as though it does. Each protected service picks the hours it's covered for, and that's what sets its price.
Business hours
For the services that matter on Tuesday afternoon and can wait until Wednesday morning — the internal dashboard, the batch job, the staging cluster. Out of hours the alert still opens an incident and still lands in Slack; what waits for the window to open is the part that wakes someone up.
Round-the-clock
For everything that has to be fixed at 3am on a Sunday whether anyone likes it or not. Full escalation the moment the alert lands, every hour of every day, through SMS and voice as well as Slack.
Nothing is ever dropped
An out-of-hours alert on a business-hours service still opens an incident, still posts to Slack and email immediately, and still sits at the top of your dashboard. Only the channels meant to wake someone — SMS, voice and mobile push — wait for the window. Cheaper cover buys a quieter night, never a blind spot.
Raise cover mid-incident
If a service you called non-critical turns out to matter at 2am, one click promotes it to round-the-clock and the escalation fires from that moment. No purchase order, no sales call, no waiting for a billing cycle. Nobody should ever have to choose between coverage and procurement.
You're billed for the cover you held
A service bills at the highest level it held during the cycle, so an emergency promotion costs you the difference for that month and nothing more. Drop it back whenever you like and the lower price applies from the next cycle.
Worked through: 12 services, 4 of which genuinely have to be fixed overnight. That's $149 + 4 × $9 + 8 × $4 = $217 a month, of which the 8 quiet ones account for $32 — about what a single seat costs on a per-user pager. Keeping the long tail that cheap is the point: you should never be better off collapsing forty real services onto one catch-all route just to trim the bill.
What we never charge for
Everything the definition leaves out. None of it moves your bill, at any volume.
Every user
Responders, managers, the other 180 engineers. Everyone gets a real account, sees incidents and reads the timeline. We will never charge you for making more of your team reachable.
Every rotation & schedule
Put three people on a rotation or fifteen — same price. Spreading on-call load so nobody burns out costs you nothing, which is the opposite of how per-seat pricing works.
Every escalation policy
Reuse one across forty services or write forty. The policy is the fire department, not the building — you're not billed for it.
Every integration
One service can ingest from Datadog, Sentry and CloudWatch at once. You pay per response path, never per alert source.
Metered, à la carte
Three things scale with what you actually consume. Everything else is flat, regardless of how hard you use it.
SMS & voice past the allowance
Every protected service contributes to one org-wide pool, sized by its cover level — a round-the-clock service brings more to the pool than a business-hours one, because it can page at any hour. Past the pool you pay $0.05 a page in the US and Canada; international is zone-rated, because a carrier bills us per message and that rate varies by destination.
Extended incident history
Alerts, incidents and the full paging timeline are kept 90 days on Pro. Keep them longer if you need to — priced by how long you keep it, because storage is what it costs us.
Extended audit retention
The audit log itself is in Pro and always will be, with 12 months of history — enough to survive an audit. This buys a longer window for teams whose compliance calendar is measured in years.
The pooled allowance per service isn't set yet — we're measuring it from real migrations rather than guessing at it, so the number reflects what on-call teams actually send. It'll be published here in full, and applied to nobody until it is. Questions in the meantime: hello@gravityloop.io.
What the price is and isn't
Three commitments that decide how this page changes over time. They're in our terms, not just our marketing.
Price tracks cost
It moves when delivering the promise gets more expensive, and for no other reason. You get 30 days' notice before it applies.
Never a charge for coverage
Adding a person, a rotation, or an escalation stage makes your on-call safer. We will never put a price on any of them.
No tier above Pro
Every capability we ship lands in Pro. There is no higher plan holding features back, and no sales call standing between you and the whole product.
Every plan, line by line
The whole thing, not a highlights reel.
| Capability | Solo | Pro |
|---|---|---|
| Coverage | ||
| Protected services | 1 | Unlimited |
| Round-the-clock cover | Included | $9 / service / month |
| Business-hours cover | — | $4 / service / month |
| Change cover level | — | Any time, self-serve |
| Users | Unlimited | Unlimited |
| Rotations & schedules | 1 | Unlimited |
| Escalation policies | 1 | Unlimited |
| Integrations per service | Unlimited | Unlimited |
| Paging | ||
| Slack, email, push & webhooks | Unlimited | Unlimited |
| SMS paging | — | Pooled, then metered |
| Voice paging with IVR ack | — | Pooled, then metered |
| Pooled pages per service | — | Included allowance |
| Platform | ||
| All named parsers | + | + |
| Schedules as code (Terraform) | — | + |
| Audit log | — | + |
| SSO & SCIM | — | + |
| Retention | ||
| Incident & alert history | 7 days | 90 days |
| Audit log retention | — | 12 months |
| Extend either, à la carte | — | + |
Pricing questions
Why charge per service instead of per user?
Because per-seat pricing on an on-call tool charges you for making your coverage better, and teams respond exactly how you'd expect — they leave people off the rotation to control cost, share a login, keep the new engineer off the schedule another quarter. For a product whose whole job is waking the right person, charging per person is charging for the thing that makes you safer. It's the same defect as an SSO tax, just harder to see. So users are free, forever, at every size.
What counts as a protected service?
An enabled service with an escalation policy attached. If it can wake someone up, it's insured and you're paying for the guarantee that it will. Integrations, users, rotations, schedules and escalation policies are all free and unlimited — one service can ingest from Datadog, Sentry and CloudWatch at once and it's still one service. You're paying per response path, not per alert source.
Doesn't cheaper daytime cover just tempt teams to under-insure?
It would if we let you get it wrong quietly. So we don't: if a business-hours service keeps generating alerts outside its window, we surface it and suggest raising the cover — the same way we surface services that have been idle for 90 days and suggest turning them off. Both directions of that nudge cost us money, which is roughly how you can tell we mean it. The alternative — one price for every service — doesn't make anyone safer, it just makes the long tail expensive enough that teams collapse forty real services onto one catch-all route and lose the ownership signal entirely.
Do I pay for a service that never paged anyone?
Yes, and that's the honest version. Coverage was held for all 720 hours of the month whether or not it was drawn on — you don't get your fire insurance back because the building didn't burn. Billing only for services that fired would mean your quiet month is cheap and your outage month is expensive, which is precisely backwards. The counterweight is that we'll tell you: any service with no alerts in 90 days gets surfaced in your dashboard with a one-click disable. We'd rather point at the four things you're paying for and may not need than quietly collect on them.
Why is SMS metered when everything else isn't?
Because it's the one thing that costs money every single time it happens. A text costs us about a cent, paid to a carrier, per message. Slack, email, push and webhooks cost effectively nothing, so they're unlimited and always will be. Every protected service contributes to one org-wide pool of SMS and voice; past the pool you pay for what you send. Charging everyone a higher flat fee to cover the loudest accounts would just be a hidden version of the same bill.
Doesn’t metering contradict "no gates"?
It would if we metered a feature. We don't. Every capability — SSO, SCIM, RBAC, schedules, escalation, the audit log — is in Pro and stays there regardless of how hard you use it. What's metered is consumption: messages sent, and how long data is stored. The rule we hold ourselves to is meter consumption, never capability.
Why isn't SSO on a higher tier?
Because it's a basic security control, not a premium feature. Charging a multiple for single sign-on is the clearest signal a vendor is pricing on leverage rather than cost. SSO, SCIM and the audit log are all in Pro.
How does retention work?
Incident and alert history is kept 90 days on Pro, and the audit log 12 months. Twelve months on the audit log is deliberate — SOC 2 and ISO audits look back a year, and a shorter window would make "the audit log ships in Pro" true on paper and useless in practice. If your compliance calendar runs longer, you extend either window à la carte, priced on how long you keep it, because storage is the actual cost. Reading your audit log is never metered. Only the length of the window is.
What happens when the price changes?
It only moves if the real cost of delivering the service moves — carrier rates, redundant infrastructure, answering support at 3am. It will never move because we found a new way to charge you. If it does change, you get at least 30 days' notice before it applies to your subscription.