Automate Web Data Extraction to CSV

Stop Copy-Pasting Website Data

Need structured web data, but hate manual scraping?

Do Browser lets you describe what you want in plain English, then handles the browser work and writes a clean CSV to /workspace.

That means you can go from "I need this data" to "I have a usable CSV" without hand-copying rows all afternoon.

Why This Workflow Works

You can:

• Extract repeating data from a page (or multiple pages)
• Keep a consistent column format
• Deduplicate with a stable key like URL or ID
• Rerun the same prompt later when data changes

Great Fit For

Use this when you need datasets like:

• Product names, prices, ratings, and URLs
• Directory listings and company metadata
• Job titles, locations, and posted dates

Prompt You Can Paste

Use this template in Do Browser:

Extract all visible [item type] from the current page.
For each row, capture: [field_1], [field_2], [field_3], [field_4].
Use [stable_field] as the dedupe key and keep the first occurrence.
If pagination exists, follow "Next" for up to [N] pages.
Write the final data to /workspace/[filename].csv with a header row.
Then share it as [filename].csv using a dobrowser link and report row count + 3 sample rows.

Example prompt:

Extract all visible products from the current page.
For each row, capture: name, price_text, rating_text, product_url.
Use product_url as the dedupe key and keep the first occurrence.
If pagination exists, follow "Next" for up to 5 pages.
Write the final data to /workspace/products.csv with a header row.
Then share it as products.csv using a dobrowser link and report row count + 3 sample rows.

How It Looks In Practice

1. Open the target website in Chrome (and sign in first if the data requires login).
2. Start Do Browser by clicking the extension icon, or press Cmd + Shift + ..
3. Paste your extraction prompt.
4. Open the generated CSV from the chat file chip (or from the Files view).
5. Preview it in the built-in viewer, then download when you are ready.

What You Get

Your output should include:

• A single header row
• One row per extracted item
• Consistent columns ready for analysis
• A file saved at /workspace/[filename].csv that you can preview and download

CSV shape example:

name,price,rating,product_url
"Product A","$29.99","4.8","https://example.com/a"
"Product B","$18.50","4.5","https://example.com/b"

Ready to Run It?

If you already have the page open in Chrome, copy the template above, swap in your fields, and run it.