servo-fetch

Node.js bindings for servo-fetch — fetch, render, and extract web content with an embedded Servo browser engine.

• No Chromium — single self-contained binary, bundled in the package
• JavaScript execution — full Servo engine with SpiderMonkey
• Schema extraction — declarative CSS-selector → structured JSON, no LLM
• Streaming crawl — for await over pages as they complete
• Typed — first-class TypeScript types, ESM and CommonJS

Install

npm install servo-fetch

The prebuilt binary for your platform is selected automatically through optionalDependencies. Supported targets: macOS (arm64, x64), Linux glibc (x64, arm64), Windows x64.

Quick Start

import { fetch, extract, crawl, map, screenshot } from "servo-fetch";

const md = await fetch("https://example.com"); // readable Markdown

const article = await extract("https://example.com"); // Readability data
console.log(article.title, article.textContent);

for await (const page of crawl("https://docs.example.com", { limit: 50 })) {
  if (page.ok) console.log(page.url, page.title);
}

const urls = await map("https://example.com"); // sitemap discovery, no render
const png = await screenshot("https://example.com", { fullPage: true });

Schema Extraction

import { extractSchema } from "servo-fetch";

const products = await extractSchema("https://shop.example.com", {
  baseSelector: ".product",
  fields: [
    { name: "title", selector: "h2", type: "text" },
    { name: "price", selector: ".price", type: "text" },
    { name: "url", selector: "a", type: "attribute", attribute: "href" },
  ],
});

CLI

The bundled binary is also runnable directly:

npx servo-fetch "https://example.com"
npx servo-fetch "https://example.com" --format png -o page.png

API

Function	Returns	CLI equivalent
fetch(url, opts?)	Promise<string> (Markdown)	default
fetchHtml(url, opts?)	Promise<string>	--format html
fetchText(url, opts?)	Promise<string>	--format text
extract(url, opts?)	Promise<Article>	--format json
extractSchema<T>(url, schema, opts?)	Promise<T>	--schema
screenshot(url, opts?)	Promise<Buffer>	--format png
evaluate(url, expr, opts?)	Promise<string>	--js
batchFetch(urls, opts?)	Promise<BatchResult[]>	parallel fetch
crawl(url, opts?)	AsyncGenerator<CrawlResult>	crawl
crawlAll(url, opts?)	Promise<CrawlResult[]>	crawl
map(url, opts?)	Promise<MappedUrl[]>	map
version()	Promise<string>	--version

Single-page calls accept { timeout, settle, userAgent, cookiesFile, headers, selector, visibility, signal }; crawl and map also accept headers.

Calls share one persistent servo-fetch process (warm engine), spawned on first use; shutdown() stops it.

Env: SERVO_FETCH_BINARY_PATH overrides binary resolution; SERVO_FETCH_ALLOW_PRIVATE=1 relaxes the SSRF guard for loopback/private addresses (trusted environments only).

Develop

Requires pnpm.

pnpm install
pnpm run build      # tsdown
pnpm test           # vitest
pnpm run typecheck  # tsc --noEmit
pnpm run lint       # biome (lint + format + import sort)