Documentation · JavaScript

Call the crawler from a typed Node service.

The library returns pages, structured failures, and final statistics so applications do not have to parse terminal output.

Browse documentation

01 · Install

Pin the package in the application.

terminal
npm install cockroach-crawler

02 · Execute

Use crawlDetailed when failures matter.

crawl.mjs
import { crawlDetailed } from "cockroach-crawler";

const controller = new AbortController();
const result = await crawlDetailed({
  seeds: ["https://example.com/docs"],
  allowedOrigins: ["https://example.com"],
  maxPages: 25,
  maxRequests: 120,
  maxDepth: 2,
  maxDurationMs: 60_000,
  maxTotalBytes: 10_000_000,
  includeSitemaps: true,
  signal: controller.signal
});

console.log(result.pages);
console.log(result.failures);
console.log(result.stats);

03 · Handle

Treat a partial crawl as an explicit state.

Inspect failures and stats before indexing. Keep page URLs and hashes beside derived chunks, and keep crawled text in a data channel because page content is untrusted.