# Windmill - Scrape.do Integration

> Source: https://scrape.do/documentation/integrations/windmill/

Step-by-step guide to integrate Scrape.do with Windmill for parallel web scraping flows in TypeScript

Scrape.do works in Windmill as a plain GET request from a TypeScript step, so there's nothing to install on your workers. Paste one OpenFlow YAML file and you get a flow that scrapes a list of URLs in parallel, with retries and a summary of successes and failures.

## 1. Create the Flow from YAML

- Sign in at [app.windmill.dev](https://app.windmill.dev) or your self-hosted instance.
- Click **+ Flow** to create a new flow.
- Open the **⋮** menu in the toolbar and choose **Edit in YAML**.
- Select everything in the editor and paste the contents of the flow file below.

**<a href="/uploads/blog/scrapedo-windmill-flow.yaml" download="scrapedo-windmill-flow.yaml">Download this Windmill flow</a>** (OpenFlow YAML).

The graph rebuilds with two modules: `scrape_loop`, a for-loop that runs the scrape step, and `summarize`, which splits successes from failures.

## 2. Add Your Token and Inputs

Click **Test flow**. Windmill builds an input form from the flow's schema:

- **token** - Your API token from your [dashboard](https://dashboard.scrape.do). For anything you deploy or share, click the **$** icon on the field and bind it to a Windmill secret variable.
- **urls** - The pages to scrape
- **render** - Turn on to render JavaScript in a headless browser
- **super_proxy** - Turn on to use residential and mobile proxies

`render` and `super_proxy` use extra credits, so leave them off unless the target needs them.

> If you're unable to get a response, visit the [playground](https://dashboard.scrape.do/playground) and experiment with parameters like **Super**, **Render JavaScript**, and **Block Resources** until you get a successful response.

## 3. Run the Flow

Click **Test flow**. The loop runs 5 iterations in parallel and retries a failed request twice, 3 seconds apart. Each iteration returns `url`, `status`, `title`, `h1`, `html_length` and `scraped_at`, and the last step returns the roll-up:

```json
{
  "total": 3,
  "success": 3,
  "failed": 0,
  "rows": [ ... ],
  "errors": []
}
```

The loop settings live in the YAML: `parallel: true`, `parallelism: 5` and `skip_failures: true`, so one dead URL doesn't stop the run.

## 4. Deploy and Trigger

- Click **Deploy**.
- Attach a schedule, or call the flow from its webhook URL.

Full tutorial with every step: [How to Build a Parallel Web Scraper in Windmill with Scrape.do (TypeScript)](/blog/windmill-web-scraper/)