| title | Checkly Documentation |
|---|---|
| sidebarTitle | Home |
| description | The active monitoring layer for engineers and agents. Detect, communicate, and resolve failures with checks defined as code. |
| mode | frame |
| canonical | https://www.checklyhq.com/docs/ |
import { CopyPromptButton } from '/snippets/copy-prompt-button.jsx';
> **For AI agents:** Start with [llms.txt](https://www.checklyhq.com/docs/llms.txt) for the full documentation index. Append `.md` to any docs URL to get its Markdown. To set Checkly up in a project, follow the [Quickstart](https://www.checklyhq.com/docs/quickstart.md).Checkly is the active monitoring layer for engineers & agents. Detect downtime with Playwright-powered monitors defined in TypeScript running from global or internal locations. Communicate outages with custom alerts and status pages. Resolve errors with traces and AI root cause analysis for new failures.
Run this in your project directory. It installs the Checkly CLI and Checkly Skills, creates a checkly.config.ts, and copies a starter prompt for your coding agent:
npx checkly initTo install only the skills for Claude Code, Cursor, Codex, and other agents:
npx checkly skills install- Detect: synthetic monitoring, uptime monitoring, and testing
- Communicate: alerts, status pages, and dashboards
- Resolve: traces, Rocky AI root cause analysis, and automatic check repair
{/* Hero */}
Guides and reference for testing, monitoring, communicating, and debugging issues in your applications.
Set up Checkly monitoring in this project.
1. Run `npx checkly init` in this project, installing the Checkly skill for this agent. Load the installed skill before writing monitoring code.
2. Inspect the project structure, existing tests, and endpoints. Create a small set of monitors as code that match this codebase: URL monitors for key pages, API checks for important endpoints, and Browser Checks or a Playwright Check Suite if a Playwright config exists. Start with the most critical flow, not full coverage.
3. Ask me for an alert email address or Slack channel and create at least one alert channel assigned to the checks. Do not use placeholder values.
4. Run `npx checkly test` and fix anything that fails. If I am not logged in, ask me to run `npx checkly login`.
5. Show me what you changed and stop for my confirmation before running `npx checkly deploy`.
Documentation index: https://www.checklyhq.com/docs/llms.txt (append .md to any docs URL for Markdown).{/* Detect / Communicate / Resolve */}
You can use Browser Checks to run plain `@playwright/test` spec in your repo on a schedule from locations around the globe, alongside API checks, URL monitors, and heartbeats.
import { test, expect } from '@playwright/test'
test('homepage loads', async ({ page }) => {
const response = await page.goto('https://www.checklyhq.com')
expect(response?.status()).toBeLessThan(400)
await expect(page).toHaveTitle(/Checkly/)
})<div className="home-tab-links">
<a className="home-tab-cta" href="/detect/synthetic-monitoring/browser-checks/overview">Create your first check <Icon icon="arrow-right" size={14} /></a>
<a href="/detect/synthetic-monitoring/api-checks/overview">API Checks</a>
<a href="/detect/uptime-monitoring/overview">Uptime Monitors</a>
<a href="/detect/synthetic-monitoring/playwright-checks/overview">Playwright Check Suites</a>
<a href="/detect/testing/overview">Testing</a>
</div>
Define alert channels once and attach them to any check or group. Route failures to Slack, email, PagerDuty, or webhooks, and publish a status page from the same project.
import { SlackAppAlertChannel } from 'checkly/constructs'
// Install the Checkly Slack App in your Slack workspace before deploying.
export const slack = new SlackAppAlertChannel('ops-slack', {
slackChannels: ['#ops'],
sendFailure: true,
sendRecovery: true,
sendDegraded: false,
})<div className="home-tab-links">
<a className="home-tab-cta" href="/communicate/alerts/overview">Set up alerting <Icon icon="arrow-right" size={14} /></a>
<a href="/communicate/alerts/channels">Alert channels</a>
<a href="/communicate/status-pages/overview">Status Pages</a>
<a href="/communicate/dashboards/overview">Dashboards</a>
<a href="/communicate/maintenance-windows/overview">Maintenance Windows</a>
</div>
Rocky AI uses your check's errors, logs, screenshots, and available traces to explain what failed, identify the likely root cause, and assess user impact. By default, Rocky AI analyzes every new error group automatically, and you can start an analysis from any failing result.
npx checkly checks get <check-id> --error-group=<error-group-id><div className="home-tab-links">
<a className="home-tab-cta" href="/ai/rocky-ai">Explore Rocky AI <Icon icon="arrow-right" size={14} /></a>
<a href="/resolve/ai-root-cause-analysis/overview">Root cause analysis</a>
<a href="/resolve/automatic-check-repair">Automatic check repair</a>
</div>
{/* Monitoring use cases */}
{/* What's new. Refresh these cards from https://www.checklyhq.com/changelog/ when a new monthly post ships. */}
Create and manage checks from your AI agent
The Checkly MCP server can now create, update, and delete checks, monitors, and check groups, so an agent can go from "this endpoint matters" to a running monitor without leaving the conversation.
Read more
Automatic check repair (Beta)
Rocky AI proposes, tests, and applies fixes for repairable errors such as outdated selectors in Browser and Multistep checks.
Read more
Status pages rebuilt around components
Group services into components with StatusPageV3, and let automation rules open and resolve incidents for you.
Read more
Pull requests from the UI
With the GitHub App installed, Export to Code and UI edits to code-managed checks open a pull request against the defining file.
Read more
Track contract usage and credit budgets
Contract customers can see consumed and remaining credits, usage trends, and projected consumption by account and check type.
Read more
Private Locations report their own errors
A new Errors tab shows error trends, stack traces, and agent details for every Private Location.
Read more
Inspect projects from the app
Monitoring as Code projects now appear in the UI with their resources, deployment history, and any drift from code.
Read more
Timezone-aware maintenance windows
Recurring maintenance windows stay at their local time through daylight saving changes.
Read more
{/* Developer tools */}
Everything you need to define, test, deploy, and automate your monitoring from code, CI, or an AI agent.
Monitoring as Code
checkly test, then ship them as always-on monitors with checkly deploy.
CLI reference
Every resource, typed and versioned
Checks, groups, alert channels, status pages, and dashboards, all defined as TypeScript constructs in your repo.
Constructs reference
Automate anything with the REST API
Manage checks, results, and account settings programmatically, with an OpenAPI spec you can generate clients from.
API reference
AI agents
Platform
{/* Connect */}