Skip to main content
Back to Solutions

You see when a release breaks a page or an API, or drops a security header

WebPixie checks your sites, APIs and staging on a schedule and alerts you when a check fails, with that check's status code and error. You can automate it with the GraphQL API and webhooks, or ask your AI assistant over MCP.

API from Starter·MCP on every plan·No credit card required

After a deploy

You shipped. Is it actually working?

A release can answer with a 200 and a blank page or an error template, and a plain ping still calls that up. An uptime monitor on your page or health endpoint passes only when the status code is one you accept and the keyword you choose is in the response, such as a build ID or "ok" in the health payload. Anything else counts as down.

API endpoints get the method, headers, JSON or form body and Basic, Token or Digest auth they need, on any port and path. Checks run from multiple global locations, such as London and New York, with an interval from 30 seconds to 24 hours depending on your plan.

Tip: your deploy script can set the new build ID as the monitor's expected keyword through the API. The next check then tells you whether the new version is the one answering.

Uptime monitors

Tagged: production

  • api.example.com/health

    GET · 200 · keyword "build 4f2c1"

    142 ms

    Up
  • app.example.com

    200 · keyword "Sign in" missing

    Down
  • api.example.com/graphql

    POST · JSON body · Token auth

    231 ms

    Up

Side effects

The release changed something you didn't touch

A framework upgrade drops your Content-Security-Policy, or a proxy change breaks the HTTPS redirect. The Main Page Analyzer checks your homepage every day, including its security headers, the redirect chain with any switch back to HTTP along the way, TLS, canonical and browser console errors. Each finding has a severity and fix guidance, and earlier daily snapshots show when it changed.

The Link Crawler follows internal links about once a week and lists pages that now return a 404 or a server error, together with the pages that still link to them.

Findings · example.com

Daily homepage analysis

  • No Content-Security-Policy header

    Security headers

    High
  • JavaScript error in the browser console

    Browser diagnostics

    Medium
  • /docs/v1 returns 404

    Linked from 6 pages

    Broken link

Alerts

You hear it before your client does, without an alert on every deploy

When a check fails, WebPixie emails your workspace on every plan, posts to Slack from Starter and sends webhooks to your own endpoint or on-call tool from Pro. From Starter, each failure becomes an incident with its duration, how often it came back and your notes, and it closes when the check passes again.

Deploys stay quiet if you want them to. Your deploy script pauses the monitor through the API and resumes it after the release, and the paused window does not lower your uptime rate. Timeouts and connection errors are retried before an incident opens, paid plans confirm them from the monitor's other selected locations, and a monitor can be set to alert only after several failed checks in a row.

Deploy script

app.webpixie.io/graphql

Before the release

mutation {
  pauseUptimeLink(ids: ["uptime_123", "uptime_124"]) {
    affectedRecords
  }
}

After the release

mutation {
  resumeUptimeLink(ids: ["uptime_123", "uptime_124"]) {
    affectedRecords
  }
}

Certificates and domains

The staging certificate should not expire before the demo

An expired certificate on staging blocks QA and the client review, and the browser greets everyone with a security warning. SSL/TLS Monitoring checks the certificate each site serves every day and warns you from 15 days before it expires, with the level rising as the date gets closer. Domain Monitoring warns you about the domain from 30 days out. If you add staging as its own site and tag it, it gets the same checks as production.

Certificates Manager lists the certificates issued for your domain and its subdomains in public Certificate Transparency logs. A preview host you forgot about shows up there, with an Add Site button to start monitoring it.

Certificates and domains

example.com and subdomains

  • staging.example.com

    SSL certificate

    6 days

    High
  • example.com

    Domain registration

    24 days

    Medium
  • api.staging.example.com

    SSL certificate

    61 days

    Valid
  • preview.example.com

    Found in Certificate Transparency logs

    Add site

Search and AI crawlers

Did a deploy hide you from Google, Bing and AI crawlers?

A staging robots.txt with "Disallow: /" or a noindex tag that reaches production throws no error. Search traffic slips over the following days, and the cause often turns up in Search Console long after the release.

The Indexability Checker reads your robots.txt and llms.txt line by line in the daily analysis. It shows which AI crawlers, such as GPTBot and ClaudeBot, have explicit rules, which Content-Signal directives are set and whether a Sitemap line is present. From Starter, Sitemap Monitoring also flags crawled pages that are missing from your sitemap.

Want an alert for this? If you add "noindex" as a keyword that must not appear on your homepage monitor, the next check fails if the tag ships.

robots.txt · example.com

Daily analysis

  1. User-agent: *
  2. Disallow: /
  3. User-agent: GPTBot
  4. Disallow: /
  5. Content-Signal: search=yes, ai-train=no

What WebPixie reports

  • Syntax

    Valid
  • AI crawlers with explicit rules

    GPTBot
  • Content-Signal

    search=yes, ai-train=no
  • Sitemap line

    Missing

GraphQL API

You can set up each new environment from code, not from the dashboard

The GraphQL API on Starter, Pro, Enterprise plans has more than 20 queries and 15 mutations. Your scripts can create, update and delete sites and uptime monitors, pause and resume monitors, tag them by environment, and read incidents, analysis findings and certificates.

API keys use Basic auth and act with the role of the user who created them. A key can be read-only and limited to specific sites and uptime monitors, so a CI job only touches the environment it deploys.

GraphQL

app.webpixie.io/graphql

Query

query {
  findUptimeLink(limit: 10) {
    items {
      id
      name
      target
      status
      checkInterval
      lastRecord {
        accessibilityStatus
        spendTime
      }
    }
  }
}

Response

{
  "data": {
    "findUptimeLink": {
      "items": [{
        "id": "uptime_123",
        "name": "WebPixie uptime",
        "target": "https://webpixie.io",
        "status": "ACTIVE",
        "checkInterval": 300,
        "lastRecord": {
          "accessibilityStatus": "UP",
          "spendTime": 0.142
        }
      }]
    }
  }
}

MCP server

You can ask your assistant what broke

The MCP server works on every plan, including Free, and connects Claude, ChatGPT, Codex and VS Code to your account through OAuth sign-in, with no key to manage. You can ask which monitors failed after the last deploy, what the latest analysis found or which certificates expire soon, without opening the dashboard.

On the consent screen you choose read-only or read and write for each workspace and can narrow it to specific sites and uptime monitors. With write access, "Pause the staging monitor while we deploy" works as a prompt.

Claude Code

WebPixie MCP · read and write

  • Did anything fail after the 14:00 deploy?
  • Yes. app.example.com has had an open incident since 14:06: it still returns a 200, but the keyword "Sign in" is missing from the page. api.example.com has passed every check since the deploy.
  • Pause the staging monitor while we roll back.
  • Done, staging.example.com is paused. Tell me when the rollback is finished and I will resume it.

You can connect WebPixie to your workflow

You can start with the dashboard, then automate what you repeat.

  1. 1

    You add your sites and endpoints

    You add the app, the API and staging, with the status codes and keyword each one should return.

  2. 2

    You choose where alerts go

    Email works on every plan, Slack from Starter and webhooks from Pro. You can turn alerts off for monitors that should stay quiet.

  3. 3

    You connect your tools

    You can connect your AI assistant over MCP on any plan, or create a scoped API key from Starter for your deploy scripts.

API access is included on Starter, Pro, Enterprise plans, and the MCP server works on every plan. You can compare plans on the pricing page and find the reference in the API docs.

Frequently Asked Questions

Common questions for developers using WebPixie.

You can point an uptime monitor at a page or health endpoint and require a keyword only the new release returns, such as a build ID, plus your accepted status codes. A plain ping counts any 200 as up, even a blank page or an error template, so the keyword is what proves the right version is answering. In uptime monitoring you can also set the HTTP method, headers, a JSON or form body and Basic, Token or Digest auth, which covers API endpoints as well as pages. Your deploy script can update the expected keyword through the API integration on Starter, Pro, Enterprise plans, so each release is checked against its own build ID. The check happens on the monitor's next scheduled run, with an interval from 30 seconds to 24 hours depending on your plan. A keyword mismatch counts as a failure on the first check, without the retries used for timeouts. The intervals for each plan are listed on the pricing page.

You can pause the monitor before the release and resume it afterwards, from your deploy script through the API or by asking your AI assistant over MCP. While a monitor is paused it is not checked and sends no alerts, and the paused window counts as unmonitored, so a planned pause does not lower your uptime rate. With the API integration on Starter, Pro, Enterprise plans, pauseUptimeLink and resumeUptimeLink take up to 100 monitors per call; the MCP server works on every plan and understands a prompt like "Pause the staging monitor while we deploy". Outside planned pauses, uptime monitoring already filters out short blips: timeouts and connection errors are retried before an incident opens, and paid plans confirm them from the monitor's other selected locations. A wrong status code or a missing keyword is a real answer from your server, so it counts on the first failed check. If a restart tends to return a few errors, you can set the monitor to alert only after several failed checks in a row.

Yes, WebPixie can monitor staging, development and preview environments the same way as production, as long as their URLs are reachable from the internet. Protected staging works too: uptime monitoring can send Basic, Token or Digest credentials or custom headers, check the status codes you accept and require a keyword in the response. For certificate and domain checks, you can add staging as its own site so it gets daily SSL monitoring and expiry warnings; each site counts toward your plan's site limit. You can tag staging, development and production monitors by environment to filter them in the dashboard and the API, and turn off notifications for monitors that should stay quiet. During a deploy, you can pause the staging monitor and resume it afterwards; the paused time does not count against its uptime rate. Keeping every environment in one workspace with tags means one plan covers them all. You can compare site and monitor limits on the pricing page.

Yes, the daily analysis reads your robots.txt and llms.txt line by line, so a "Disallow: /" line or a removed AI crawler rule shows up in the next analysis. The Indexability Checker shows whether the files are valid, which AI crawlers such as GPTBot, ClaudeBot, Google-Extended and PerplexityBot have explicit rules, which Content-Signal directives are set and whether robots.txt has a Sitemap line. The Main Page view also shows your homepage's canonical URL and whether it serves Markdown to clients that ask for it. From Starter, Sitemap Monitoring flags crawled pages that are missing from your sitemap. These results appear in the analysis rather than as an alert. If you want an alert, you can add "noindex" as a keyword that must not appear on your homepage uptime monitor, and the next check fails if a noindex tag ships to production.

Yes, your pipeline can call the GraphQL API to create, update, pause and resume monitors and to read monitor state, incidents and analysis findings after a deploy. A common setup pauses the production monitor before the release, resumes it afterwards and sets the expected keyword to the new build ID, so the next scheduled check confirms the right version is live. The API integration is available on Starter, Pro, Enterprise plans and uses API keys over Basic auth. A key can be read-only and limited to specific sites and monitors, so a CI job only touches the environment it deploys. Monitor creation is also capped per day at your plan's monitor limit, so it works best to keep one monitor per environment and update its target instead of creating a new one for every preview deploy. Checks run on each monitor's own interval. To get alerts into your own tooling, webhooks send incident events to an HTTPS endpoint you set, and each event matches a record in incident management. Webhook availability depends on plan, and you can compare plans on the pricing page.

Yes, WebPixie has a GraphQL API with more than 20 queries and 15 mutations, included on Starter, Pro, Enterprise plans. The API integration covers sites, uptime monitors, incidents, tags and workspaces, so a deploy script can create a monitor for a new environment, pause it during a release, resume it afterwards and read the latest analysis findings. Authentication uses API keys over Basic auth. A key acts with the workspace role of the user who created it, can be read-only and can be narrowed to specific sites and uptime monitors, so you never have to share a broad workspace credential. Teams also pull uptime history into internal dashboards and sync incident data with their own tools, and for alert-driven workflows the API pairs well with incident management and webhook notifications. The API docs cover endpoints and authentication setup. If you would rather work through an AI assistant than code, the MCP server connects Claude, ChatGPT, Codex and VS Code to your account on every plan.

Everything you can monitor through WebPixie

Every check is available in the dashboard, and through the API on Starter, Pro, Enterprise plans.

Related Solutions

You can start monitoring your sites and APIs today

Free plan with MCP, no credit card required.