Give it a page to start from and the words you care about. It reads the open web one page at a time and brings back every page that mentions them. Every crawler lives here, in one world that grows with every site it reads.
Free. Runs entirely in your browser — nothing is uploaded.
MY CRAWLER
Name it, pick where it starts, and tell it which words matter. It walks the open web from there — fetching real pages, following real links — and every page it reads lands in Finds and lights up a town on the map.
FINDS
Every row is a page that was actually fetched and read in this browser. The score is how many times your words appear in it. Click a row to open the real page.
TOWNS
One town per domain. A town stays dark until a crawler actually reads a page there — then its windows come on and it keeps the pages. Founding towns sit on the map from the start, unlit, waiting for someone to visit.
HOW IT WORKS
No servers, no queue, no account. The crawler is a loop running in this tab, and everything on the map comes out of what it really fetched.
Either an encyclopedia article you name, or the live Hacker News front page. Both of those serve pages that a browser is allowed to read across origins, which is what makes a crawler possible without a backend.
It pulls the page text, counts how often each of your words appears, and records the title, the URL and the score. A page that matches nothing is still recorded — knowing where it looked matters as much as what it found.
Every link out of the page goes into a queue. It takes them in order, skipping anything it has already read, until it hits the page limit you set or you stop it.
Every domain it touches becomes a town. Read a page there and the town lights up and keeps the page. The map is nothing more than a drawing of what your crawlers have actually seen.
Browsers refuse to hand JavaScript the contents of another site unless that site says it is allowed. Most of the web says no. A crawler built this way reads the part of the web that says yes — and writes down a refusal for everything else rather than pretending it got in.
In this browser. Crawlers, finds and towns are kept in local storage on your own machine. Nothing is uploaded, there is no account, and clearing it is one button on the crawler page.
The world is generated once from a fixed seed, so it is the same map every time you open it. Towns sit at a position derived from their domain name, which is why a domain always lands in the same place. Crypto domains settle on the coast.
It is the coin attached to this. The world, the crawler and every find work with no wallet connected and no coin held — that is the point. The token page carries the contract and the live chart.
$KRAWD
Read live from the pair every 45 seconds. Anything the read does not return is left as a dash.
the line starts when the first reading lands
Phantom, Solflare or Backpack. Keep the seed phrase on paper and off every screen.
From an exchange or the wallet's own on-ramp. Leave a little spare for fees.
Copy it above. Never from a reply or a screenshot — there will be fakes.