A Guide to Custom Monitoring Criteria with Natural Language

Why Plain English Beats CSS Selectors for Monitoring

Setting up a traditional web monitor used to mean picking a CSS selector, hoping the class name didn't change, and writing a script to compare the result week over week. The first time the site's front-end team renamed .price-tag to .product__price, the monitor went silent without telling you. WebMonitor.fyi lets you describe what you want to track in plain English instead — and the AI handles the rest, even when the underlying HTML changes. This guide covers how to write good natural-language monitoring criteria, from basic patterns to combined-condition rules.

Why Natural Language Criteria Work

The case for natural-language queries is covered in this in-depth article from Alation. Three operational gains for monitoring specifically:

• No technical setup. Describe what you want to track in a sentence. No selectors, no scripts, no maintenance when the site's markup churns.
• Resilience to markup changes. The AI reads the page semantically. When a site refactors its CSS or moves a price into a new component, the monitor keeps working.
• Accessibility across the team. Anyone — not just developers — can set up a monitor that catches what matters to their work.

What natural-language criteria don't fix: vague criteria still produce vague alerts. "Tell me when something changes" matches everything. The work moved from selector-writing to criterion-writing.

How to Set Up Your First Custom Monitor

Three steps from URL to active monitor:

1. Enter the URL. Paste the page URL into the "Add New Monitor" form.
2. Write your criteria. Describe in plain English what you want the AI to watch for. Examples in the next section.
3. Set frequency and alerts. Pick the polling cadence and notification channels (email, SMS, webhook).

Examples of Custom Monitoring Criteria

Basic Examples

• Price tracking: "Notify me when the price drops below $500."
• Stock monitoring: "Alert me if the product is back in stock."
• Keyword spotting: "Inform me if the phrase 'data breach' appears on this page."

Advanced Examples

• Sentiment analysis: "Alert me if there are any new negative reviews about our product."
• Job postings: "Let me know when a new job posting for a 'Software Engineer' is published."
• Combining criteria: "Notify me if the price of the 4K TV is below $1000 AND it is in stock."

For more applications, see our article on the top use cases for web content monitoring.

Best Practices for Writing Effective Criteria

Three rules that produce reliable monitors:

• Be specific. "Notify me if the 'Features' section of this page is updated" beats "tell me about changes."
• Use quotes for exact phrases. "Alert me if the text 'Terms and Conditions have been updated' is present" — quotes anchor the AI on the exact string.
• Start simple, then refine. A direct one-condition criterion is the right starting point. Combine conditions later once you see what the monitor catches and misses.

Set Up Your First Custom Monitor

Natural-language criteria put precise monitoring within reach of anyone on your team. WebMonitor.fyi handles the AI interpretation and page polling so you can describe what matters in a sentence and walk away. Sign up for a free account and run your first custom monitor in under 5 minutes. The pricing page lists paid plans by check frequency and monitor count.