View live status and subscribe to incidents →

Our status page is hosted independently of our own infrastructure, deliberately: a status page that runs on the systems it reports on goes down with them.

What is monitored

The status page groups its checks; this is what each one covers.

GroupComponentWhat it means
APIInference APIPOST /v1/chat/completions and the native invoke plane. This is the data plane our uptime target refers to.
APIBatch API/v1/files and /v1/batches, including whether submitted batches are being polled and completed.
APIWebhook deliveryWhether batch results and lifecycle events are reaching registered endpoints.
APIkapualabs.comThe marketing site.
Control PlaneManagement APIThe control plane’s health endpoint — the surface your API keys, provider connections and webhook registrations are managed through.
WebCustomer portalThe signed-in dashboard at /portal.
Webllm-bench.kapualabs.comThis documentation and benchmark site.

The Inference API check is a real inference call, not a ping. It asserts that a model actually answered, because an API can return a 200 with an empty body — and an availability check cannot tell the difference.

What is not on that list is not monitored, and we would rather say so than list a component nobody watches: a status page reporting green for something it does not measure is worse than one that does not mention it.

Availability and support

The table below is reproduced verbatim from §15 of our Terms of Service , which is the authoritative version.

We publish a 99.5% monthly uptime target for the data plane.

This is a target, not a warranty. Version 1 of the Service carries no service credits and no financial remedy for missing it. Where our sales or pricing materials use the word “SLA”, it refers to this target and to the support commitments in the table below, and to nothing else.

TierChannelFirst responseUptimeStatus page
FreeDocs and email, best effortnonenonepublic
Shared-anonEmailNext business dayPublished 99.5% monthly targetpublic
PrivatePriority email, named contact6 hoursPublished 99.5% monthly target, plus incident communications and postmortemspublic, plus incident subscription

First-response times are measured during support hours: 08:00–20:00 Pacific Time, Monday to Friday, excluding US public holidays. A request that arrives outside those hours starts its clock when they next open. “Next business day” means the next such day.

Reporting a problem

If you are seeing errors that the status page does not reflect, tell us — that gap is itself useful information. Include the x-llmbench-diagnostic header value from a failing response if you have one. Never send us an API key, and never paste a request or response body that might contain one.