Pull structured JSON data from any website or SPA using headless browser automation. Replace manual data entry and one-off scraping contractors. Runs 24/7, handles JS-rendered pages.

This is the whole blueprint, nothing held back: the model, the tools, the exact prompt, the source on disk, and the setup. Everything you need to run Data Extractor yourself is on this page, and it deploys in an afternoon.

What it does

Data Extractor handles the work end to end:

Extract structured data from any URL including SPAs
Handle JavaScript-rendered pages with headless browser
Return clean JSON output with configurable schemas
Support batch extraction from multiple pages

The anatomy

An agent is a composition, not a prompt: a model, the tools it is allowed to call, the integrations that give those tools reach, and the rules that keep it honest. Here is Data Extractor on one card.

Data Extractorblueprint

Model: claude-sonnet-4-6
Integrations: Browser Use
Guardrails: user gives url + what to extract; you run a browser use task on that page

The instructions

Every run is governed by this prompt. It is short on purpose: enough to make Data Extractor reliable, not so much that it can misread itself. Copy it as a starting point for any agent you build.

Data Extractor

You are a web scraping agent. You MUST use Browser Use Cloud to extract structured data from web pages. User flow: - User gives URL + what to extract - You run a Browser Use task on that page - You return clean structured JSON (url, request, data, notes) Rules: - ALWAYS use Browser Use for extraction. Never try to fetch pages directly. - Return structured JSON output matching the schema. - If extraction fails, explain what went wrong and suggest fixes. - Handle dynamic pages and SPAs that require JavaScript rendering.

The source, on disk

Three files carry the whole agent: the definition with its tools, the environment it expects, and the setup. Walk them.

data-extractor/data-extractor.ts0.5 KB

1import { agent } from "@agent-sdk"

2import { webScraperTools } from "./lib/tools"

4export default agent({

5 model: "claude-sonnet-4-6",

6 runtime: "claude-code",

7 permissionMode: "bypassPermissions",

8 maxTurns: 20,

9 systemPrompt: `...`, // see System Prompt section above

10 tools: webScraperTools,

11 onError: async ({ error }) => {

12 console.error("[agent] error:", error)

13 },

14 onFinish: async ({ cost, duration, turns }) => {

15 console.log(`Done: ${turns} turns, ${duration}ms, $${cost.toFixed(4)}`)

16 },

17})

The business case

If you run an agency, Data Extractor is also a product. It sells into marketplaces, price monitoring, research firms, and the numbers work at very small scale.

$1,800

setup you can charge

$300/mo

monthly retainer

jobs off your plate

Ship it

The whole install is a handful of commands and the keys from the env file. Run the list, then point Data Extractor at real work.

Deploy checklist

0 of 5

That is the entire blueprint: one model, a clean toolset, and a prompt. Built by hand it costs you an afternoon. Inside Ultron it costs you a sentence, because you can build this exact agent in the workforce and watch it run before you commit to anything.

Build Data Extractor in Ultron

Describe what you want

Start from this blueprint or your own words. Ultron reads the intent and picks the tools.

It builds the agent

The model, the tools, the prompt, the console. Assembled in front of you, nothing to wire.

Watch it run

Test it live in the chat, then put it to work. You review what it does, it does the rest.