Agent on-call · early access

Your agent takes the first look. You stay in control.

When something breaks, Pulsefast pages your team in seconds and starts your coding agent on the incident, with the failing checks, the error and the deploy that caused it. By the time you open your laptop, the cause is written up.

Get early access How it works

From alert to cause, before you're awake

1. An alert fires

From Sentry, Grafana, Datadog, CloudWatch, any webhook, or Pulsefast's own 5-second checks. Your team is paged right away, through your usual pager or chat. The agent never delays the page.

2. Your agent investigates

It starts on your repo at the deployed commit, with the evidence: which checks failed and where, response times, the error, and what was deployed minutes before.

3. You decide

It writes up the cause and proposes a fix as a pull request. Restarts, rollbacks and anything risky wait for a human to approve.

Plugs into what you run

Alerts from

Your agent works with

People are paged through

An agent on call, without losing control

Humans paged in seconds

Every incident reaches your team at full volume from the start. The agent is extra help, never a reason to wait.

Approvals that hold

Risky actions pause until someone approves from the incident page or the alert. Deny, and the agent adjusts.

A record of everything

Every alert, page, finding, action and approval, with timestamps. Who approved what is never a mystery.

Bring your own agent

Run Claude on your repo with your own Anthropic key, or send incidents to any agent you host, by webhook or a small listener.

Keep your pager

Pulsefast works alongside Opsgenie, PagerDuty or incident.io. Nothing to rip out on day one.

Flat price, whole team

One price per team, not per person. Everyone who should see an incident can.

Moving off Opsgenie?

Atlassian ends support for Opsgenie on April 5, 2027. If you're choosing what comes next, join early access and tell us how your team pages today. We're building the move with the first teams.

Pricing

Early-access teams get it free while we build it with them.

Team

$99/mo

  • Alerts from 10 sources: Grafana, Sentry, Datadog, CloudWatch, any JSON
  • Your agent on every incident
  • Approvals and a full record
  • 100 uptime monitors at 5-second checks
  • Whole team included

Business

$299/mo

  • Everything in Team
  • 50 alert sources, 20 agents
  • 250 uptime monitors
  • Priority support

Get early access

We're onboarding a small group of teams by hand. No spam, no sharing your address.

Questions

Does the agent touch production?

Only what you allow. By default it reads your code, investigates and proposes fixes as pull requests. Anything that changes a running system waits for a human's approval, and production access is off unless you turn it on per service.

What if the agent is wrong?

Then you've lost nothing: your team was paged at full volume from the first second, as today. The agent's findings are notes for you to check, and the incident only closes when your monitoring sees recovery, not when the agent says so.

Which agents work?

Claude, run on your repo with your own Anthropic API key, so the work is billed to and stays in your account. Or any agent you host: Pulsefast sends it the incident by signed webhook or to a small listener on any machine, and it reports back.

Do I have to change my monitoring or pager?

No. Point your existing alerts at Pulsefast and keep paging the way you do now. If you want fast uptime checks too, they're built in.

When can I use it?

Early access opens to a small group of teams first, set up with each of them. Join and tell us how your team handles incidents today.