# Introduction > What Firehose is, who it's for, and how the real-time pipeline turns your rules into a live stream of matching web pages — plus where to go next. Firehose is a real-time web monitoring platform. Instead of polling websites or scraping search results, you declare **what** you care about once — as a query — and Firehose delivers **every** crawled page that matches, as it happens. ## What you can do with it - **Track brand and competitor mentions** the moment they're published anywhere on the web. - **Monitor news and topics** by category, language, and recency. - **Detect new pages** matching a pattern (new product pages, job listings, press releases). - **Watch specific URLs** for content changes and capture the diff. ## How it works Firehose sits on top of a continuous web crawl. The pipeline is: 1. You create **rules** (queries) on a **tap** (an API token). 2. Firehose evaluates every freshly crawled page against all active rules. 3. Matches are pushed onto a stream and fanned out to your open **SSE** connection. ```text You create rules ──▶ Firehose evaluates every crawled page ──▶ Matches stream to you (SSE) ``` A tap is built for **web-scale mention tracking**: it surfaces a page when the crawler reaches it, on the crawler's own schedule. To monitor a specific list of URLs you already know, use **[URL Watch](/url-watch/overview)** instead — [Stream vs URL Watch](/get-started/stream-vs-url-watch) explains which one fits your use case. ## Where to go next Create a key, a tap, a rule, and open your first stream. The objects you'll work with and how they relate.