scrapiq-node

Dependency-free TypeScript client for Scrapiq — the open-source API that turns any URL into clean markdown or structured JSON for LLM/RAG pipelines.

Zero runtime dependencies (uses the global fetch, Node 18+). Ships ESM + TypeScript types.

Install

Not on npm yet, so install from this repo:

npm install github:NG-PR0JECT/scrapiq-node

The build runs on install (prepare -> tsc), so dist/ exists by the time your import resolves.

Quick start

import { Scrapiq } from "scrapiq-node";

const client = new Scrapiq(); // defaults to https://scrapiq.io

// Clean markdown for RAG ingestion
const page = await client.markdown("https://en.wikipedia.org/wiki/Large_language_model");
console.log(page.content); // clean markdown, no nav/boilerplate

// Structured JSON against a schema
const product = await client.extract("https://example.com/products/1", {
  format: "json",
  schema: {
    title: { type: "string" },
    price: { type: "number" },
    description: { type: "string" },
  },
});
console.log(product.data);

API

new Scrapiq(options?)

Option	Default	Description
baseUrl	https://scrapiq.io	API base URL (use your self-hosted instance)
apiKey	—	Reserved for future authenticated plans
timeoutMs	30000	Request timeout

client.extract(url, options?)

Option	Default	Description
format	"json"	"json" | "text" | "markdown"
schema	—	JSON schema for structured extraction (format="json")
includeMetadata	true	Include page metadata (title, description, author…)
signal	—	AbortSignal for cancellation

Also: client.text(url) and client.markdown(url) convenience wrappers, and client.health().

Errors are thrown as ScrapiqError with status and message from the API.

Self-hosting

Point the client at your own instance:

const client = new Scrapiq({ baseUrl: "http://localhost:8001" });

docker run -d -p 8001:8001 ghcr.io/ng-pr0ject/scrapiq  # see main repo for the image

License

MIT