Part of Project Helena

Monitor every client
from one dashboard.

Warden gives each of your clients a status page with their own branding — and their own login to watch it. You keep one dashboard for all of them.

Book a setup call Self-host it free

One install

  • A group per client unlimited
  • A status page per group branded
  • A login per client theirs only
  • Checks as fast as 10s

Billed per client, never per seat or per check.

Onboarding a client takes four steps.

Not four weeks. This is the whole workflow — there is no step five.

  1. 1.0

    Add the client

    A group per client. Every monitor you create lives in exactly one of them, so a client’s services never bleed into another’s view.

    Groups · unlimited

  2. 2.0

    Point their endpoints at it

    HTTP and HTTPS checks from 10 seconds up. Set the method, headers, body, timeout, retries and which status codes count as healthy — per monitor, not per account.

    Intervals · 10s / 30s / 1m / 5m / 10m

  3. 3.0

    Put their brand on the status page

    Each group gets its own page: own slug, logo, favicon, accent colour, light or dark theme, header layout, and an uptime window from 7 to 365 days.

    Status page · one per group

  4. 4.0

    Hand them the login

    Give the client an account that opens straight onto their page and nothing else — not your dashboard, not another client’s. The page can stay private.

    Role · status_viewer

Repeat for the next client. The install does not care whether you have three or fifty.

AI-operated uptime

Ask your AI what is down.

Warden exposes live monitors, incidents, failed checks, latency, certificates and notification policy through a role-scoped MCP connection. Your assistant can explain what happened without another dashboard hunt.

Start with a viewer key. Give a separate editor key only when the assistant should create monitors, organize groups or pause checks.

See how AI control works
  1. 01 “What is down right now?”
  2. 02 “Why did checkout fail overnight?”
  3. 03 “Which certificates expire next?”
  4. 04 “Create checks for these client domains.”

MCP endpoint · /api/mcp · bearer authentication

The parts that matter at 3am.

Alerts that don’t cry wolf

A blip is not an alert. Warden confirms failures, waits for an outage to persist, groups related failures into one incident, follows a reminder ladder, confirms recovery, and quiets a flapping monitor until it settles. Confirmation and latency thresholds can be set per monitor.

The cert nobody was watching
SSL expiry warnings at 30, 14, 7 and 1 days, and it notices when you renew. This is the outage that makes an agency look worst.
You see what actually broke
The failing check keeps the status code, the latency, the error, the response headers and the body. You open the event already knowing.
Tell them before they tell you
Publish an incident with a timeline your client can follow, promote a detected outage into one, and schedule maintenance windows that suppress alerts.
One message a day
A digest at a time you pick, to Slack or a webhook — including on the quiet days, so silence never means the monitoring died.
Your whole team, always
Unlimited members on every plan. Four roles: admin, editor, viewer, and a client-only role that sees one status page.
Ask your AI what happened
A role-scoped MCP endpoint lets your assistant inspect monitors, incidents, check history, latency and expiring certificates. An editor key can create and organize monitors without getting delete access.

Priced per client. Not per seat.

You bill your clients per client. Your monitoring should work the same way. At around $5 a client, most agencies bill it back at $25–50 a month as “monitoring and status page” — it stops being a cost and becomes a line item. Every plan includes unlimited team members.

There are cheaper uptime checkers, and we’ll name them on the call. What they don’t give you is a real account per client — one you can hand over, scope to their services only, and take back when the contract ends. Most of them hand out a shared page password and call it access.

MOST AGENCIES

Agency

$ 99 /month

Up to 20 clients

$5 per client / month

  • Unlimited monitors
  • A branded status page per client
  • A private login per client
  • Slack, webhook and email alerts
  • SSL expiry alerts at 30/14/7/1 days
  • Sustained alerts, correlation and flap detection
  • Incidents and maintenance windows
  • REST API and Google SSO
  • Unlimited team members
  • Support over Slack
Book a setup call

Pro

$ 249 /month

Up to 50 clients

$5 per client / month

  • Everything in Agency
  • Priority support, same-day
  • We migrate you off your current tool
  • Quarterly review call
Book a setup call

Enterprise

More than 50 clients, or you need it running inside your own network.

Talk to us
$ 299 once

Setup

Every paid plan starts here. We sit down with your client list, build a group and a status page for each one, put their branding on it, wire your alerts into Slack, create their logins, and move you off whatever you’re using now. You show up to the call with a list of clients and leave with it running. Hosted, updated and watched by us from then on — there is no server for you to keep alive.

$ 0

Community

Unlimited clients

Warden is AGPL-3.0 and the whole thing is on GitHub. Run it on your own box and you pay nothing, forever — same groups, same per-client status pages, same client logins, same 10-second checks. One Docker command and it is up. You keep it running, you keep it updated, and the community answers your questions.

The paid plans are not a better Warden. They are the same Warden, with us running it.

Self-host it free Capacity planning guide →

What Warden can’t do yet.

Published so you can decide with the real picture. If one of these is a dealbreaker, say so on the call and we’ll tell you straight whether to wait.

  • Checks from more than one place

    Planned

    Today every check runs from the single host Warden lives on. Remote probes are the next big piece of work.

  • Monthly SLA reports

    Planned

    A per-client uptime report you can send when the contract asks for one. Today you get the number on the status page.

  • Per-client alert routing

    Planned

    Right now every alert goes to every channel you enable. Routing a client’s alerts to their own Slack is coming.

  • Status pages on your client’s domain

    Planned

    Pages live at your Warden host today. CNAME support is planned.

How many clients are you watching?

Book a 15-minute call and we’ll set Warden up around your client list — groups, status pages, branding, logins — and show you the result. If it doesn’t fit how you work, we’ll say so.

Book a setup call

Not ready to talk? Leave your email and we’ll reach out.

No spam. One reply from a person.