What is PageLantern?
PageLantern is an external monitoring service for websites, APIs, SSL certificates, domains, and DNS. It checks your systems from the outside — there is nothing to install on your servers — opens incidents automatically when checks fail, alerts your team, and publishes public status pages.
How do I create my first monitor?
Sign up and choose Add website to open guided setup. Enter a URL you are authorized to monitor, choose your checks, and review the monitor slots before creating them. Eligible checks depend on your site and plan: SSL requires HTTPS and DNS requires a paid plan. Advanced setup and starter templates remain available for custom monitors.
By default the alert policy opens an incident after 2 consecutive failures and resolves it after the first successful check — you can change both the failure and the recovery thresholds, plus channels and measurement intervals, later.
Do I have to use AI recommendations?
No. Choose manual setup to configure checks without any AI request. If recommendations are enabled for your workspace, you can explicitly request them, review the evidence and suggested checks, and change the selection before creating monitors. An AI request alone creates no monitors.
If recommendations are disabled, unavailable, or at their request limit, continue with manual setup. Existing monitors keep running independently of the AI provider.
What is sent to the AI provider?
Only after you explicitly request recommendations, PageLantern sends OpenAI validated public target and candidate URLs, public page and sitemap links, sanitized observations, limited public page text, and supported check candidates. OpenAI can use those links or search the website domain for additional public pages. AI setup rejects URLs containing query strings or embedded credentials and does not send your configured authentication headers, cookies, or account details. Do not submit private pages or sensitive information; use advanced manual setup when your monitor needs authentication or a query string.
Suggested pages include links to where they were found: a public page, a sitemap, or public search. If PageLantern could not verify a page, its evidence says so and includes any observed HTTP status. A link or search result does not establish live availability. Review the suggested pages and checks before starting. Test & Start previews the main target; each additional page is checked by its monitor after creation. A blocked or failing HTTP response is still a monitoring failure.
Temporary sanitized recommendation drafts expire after 24 hours and are deleted by hourly cleanup. Applying a draft deletes it immediately; an encrypted retry receipt containing monitor IDs and a request hash is kept for 30 days. Accepted reasons, evidence, and public source links stay with the monitor. Raw fetched pages and prompt bodies are not written to PageLantern application or proxy logs. OpenAI stores recommendation inputs and generated outputs for authorized troubleshooting, including submitted public URLs and observations and hosted web-search activity. Standard Responses storage retains response data for at least 30 days. API data is not used for training unless the API customer opts into sharing; separate abuse-monitoring logs may retain content for up to 30 days, with documented exceptions. See the Privacy Notice and Sub-Processors page for details.
How often does PageLantern check my site?
Checks run every 5 minutes on the Free plan, every 1 minute on Starter, and as fast as every 30 seconds for critical checks on Team.
How will I know when something breaks?
PageLantern sends alerts to Email, Slack, Microsoft Teams, Discord, Webhooks, PagerDuty, and SMS. Incidents track the outage lifecycle — open, acknowledge, resolve, reopen — and can be published to your public status page.
SMS is metered rather than unlimited: the Starter plan includes 50 messages a month and the Team plan 200. Every other channel counts against your plan’s channel limit — 1 on Free, 5 on Starter, unlimited on Team.