Your incidents never leave.Nothing changes without a human.
3AM runs beside your systems. The only thing that reaches us is a signed licence check, and air-gapped, not even that.
Connected, or fully air-gappedNothing listens, either way
Connected, it reaches out only to your tools and a signed licence check. Air-gapped, it reaches out to nothing at all.
Air-gapped: the licence verifies offline, the models ship in the bundle, approvals happen in the console.
Secrets are gonebefore anything is written
Passwords, tokens, keys and card numbers are redacted on the way in. Prompts are kept redacted, with a hash for proof.
- No incident data sent to us
- Models on your GPU, or an endpoint you choose
- Card numbers caught by checksum
Humans hold the keys.Always.
Every service starts in shadow. Letting 3AM act takes two people. Taking it back takes one click.
One switch stops everything.
A record that provesit hasn't been touched
Every signal, check, approval and action, hash-chained and signed. Change one byte and the verifier shows where.
Linking…
Documents can suggest.Only evidence can act.
Every source carries a trust tier. A runbook can point at a cause; only a live check can authorise a fix.
- T1Code and schemaCan act
- T2Alert rules and live metricsCan act
- T3Runbooks and wiki pagesCan suggest
- T4Past postmortemsCan suggest
Built to be checked
- Signed, distroless images
- SBOM with every release
- Signed licence, verified offline
- Source-free compiled build
- Read-only diagnostic credentials
- No shell access for the agent
What your security team will ask
Does any of our data leave the network?
No incident data does. 3AM runs on your hardware and reaches out only to the tools you connect, such as Slack for an approval, to a model endpoint if you use one instead of a GPU, and to Warden, our licence service. About once an hour, that check sends the licence, the install's fingerprint, its version and a signed count of services and investigations. Set it offline and it never calls Warden.
Do we need to open inbound firewall ports?
No. Slack uses Socket Mode, Symphony its datafeed, and PagerDuty and other channels are polled. Every connection is opened from inside your network.
Can 3AM change production without a person approving it?
Not unless you allow it. Every service starts in shadow mode, where 3AM only proposes. At L1 each action needs one approval, raising a service needs two approvers, and a global halt stops everything at once.
What stops it acting on the wrong root cause?
Only a cause confirmed by a read-only check against your live systems can lead to a fix. Runbooks and past incidents can suggest causes but never authorise action, and every fix is dry-run first.
Which AI models does it use?
Open-weight models. With a GPU they run inside your install, and nothing calls out. Without one, you point 3AM at a model endpoint: your own, inside your network, or a hosted one. It won't run them on a CPU, where every investigation would wait.
How do our auditors verify what happened?
Every event is written to an append-only, hash-chained ledger with signed segments. An independent verifier detects any edit, deletion, insertion or reordering, and the ledger exports to your SIEM.
Can it run in an air-gapped environment?
Yes. The air-gap bundle ships the images and model weights together, so nothing needs to be fetched at install time.
Have a vendor security questionnaire? Bring it to the demo. We'll walk your team through it line by line. Book a demo ›
Your next incident is already scheduled.
See 3AM resolve one live in 30 minutes, then start a shadow-mode pilot that changes nothing.