How does Google Alerts work? The pipeline behind the email
A stored query, Google Search, a few filters and an email. What that design is good for, and what developers use when they need webhooks.
By ClaudeGPT Si Newsroom, Editorial team
3 min read
How does Google Alerts work? You save a search query on google.com/alerts, and when Google Search finds new results that match it, Google emails them to you at the pace you picked. That is the whole system: a stored query, a set of filters and an email. Knowing that tells you what it is good for and why it gets noisy, and why developers who want alerts in code or in chat often reach for AyeWatch instead.
The pipeline, step by step
- You store a query. Whatever you type in the box at Google Alerts becomes the alert. The page previews recent results for it, so you can see roughly what the alert will send before you save.
- Google Search finds new matches. In the words of Google's Create an alert help page, you get emails "whenever we find matching search results". The source of truth is Google's own search, not a crawl of pages you pick.
- Filters narrow the set. Your options decide which matches count: the type of source (automatic, news, blogs, web, video, books, discussions or finance), the language, the region, and whether you want only the best results or all of them.
- Delivery batches them. The frequency setting decides when mail goes out: as it happens, at most once a day, or at most once a week. The email goes to the Google account that owns the alert.
- You manage it in one list. Edit changes the query or options, Delete removes the alert, and every email carries an Unsubscribe link.
What that design means in practice
| Design choice | What follows from it |
|---|---|
| The trigger is a search query | Matching is about words; you cannot tell it what a result should mean |
| New results come from Google Search | You hear about new pages, not edits to a page you already know |
| Filters are source, language and region | You can narrow where results come from, not what they say |
| Delivery is email | Getting results into Slack or a script means extra plumbing |
For a developer, the last row is the one that hurts. Google's help pages describe email delivery and document no API or webhook for alerts, so wiring them into a team channel or a pipeline means parsing mail or adding a third-party tool.
Where it works well
The design is a good fit for low-volume, exact terms: the name of your open source project, a rare library, a CVE identifier you are waiting on, or a product code name. A quoted phrase plus a minus term or two keeps the volume down, and Google documents those operators for Search in Refine Google searches. It costs nothing and takes a minute.
A different model: rules, pages and webhooks
AyeWatch works from a different starting point. Instead of a query you write a rule in plain English, such as "new releases or breaking changes in this framework, ignore tutorials and opinion posts", and AI reads each new item against it, so a match means the item fits what you asked for, not only that it shares a word. Matches arrive as a short summary with links to the sources.
It also watches things Google Alerts does not: a specific URL such as a changelog, docs page or status page, and public accounts on X, Reddit, YouTube and LinkedIn, among others. On delivery, the feed and push notifications come with every plan, while email, Slack, Discord, webhooks and the REST API come with paid plans. The API lets you create, list, update and delete topics, and webhooks carry signed update events with the AI summary; the webhook docs show the format. There is a free preview, and paid plans start at $9 a month, billed in monitoring credits.
If all you need is an email when a rare string appears on the web, Google Alerts is free and good enough. Once you want the alert to understand the topic, watch a page, or reach your tools without an inbox in between, AyeWatch is the one to set up.