Pull products, prices, reviews, images and contacts out of any site with a single click — or hand the job to your AI agent and get back a finished table.
Same engine underneath — pick the workflow that fits the job.
Hover a list on the page, click it, and every row lands in an editable table in Chrome's side panel — ready to export.
Connect Claude, Cursor or any MCP-capable app once. After that, just ask in plain English — the scrape runs and a table comes back.
Hand a scrape to cloud browsers. It runs at volume, on a schedule, from the country you pick — and delivers fresh data on its own.
Every column is named and typed on its own — images, prices, links, you name it. Fix a cell, drop a column, then export to CSV, Excel, JSON or Google Sheets.
| # | Product | Category | Price |
|---|---|---|---|
| 1 | Trail Runner GTX | Running shoes | $149.00 |
| 2 | City commuter bike | Cycling | $899.00 |
| 3 | Ceramic pour-over set | Kitchen | $42.50 |
| 4 | Nordic oak desk lamp | Home office | $76.00 |
| 5 | Featherlight down jacket | Outerwear | $210.00 |
Connect once over MCP and scrape whole sites, query the results and schedule jobs — right from the conversation. Ask for “every product on this store, sent to my Sheet”, and a finished table is what comes back.
The scrape runs on our infrastructure, not in your chat window — so raw HTML never burns your context window, and huge jobs still cost a couple of sentences.
Name, price, SKU, availability and images — for every product URL on a site.
Shopify, WooCommerce, Magento — every product and variant in one export.
“Every product page on this domain” — it maps the site and picks the slice you meant.
Emails, phone numbers and social profiles across thousands of company sites.
Name, rating, address, phone and opening hours for a list of places.
Filters, sums and counts run server-side — the rows never enter the chat.
“Drop empty columns, merge address parts, remove duplicates.” Done.
“Re-run yesterday's price scrape, and every morning from now on.”
A download link, a fresh Google Sheet, CSV, Excel or JSON.
Pages render for real — JavaScript runs, lazy content loads, and consent walls get handled on the way in.
Keep a Google Sheet current on every run, or push results to a webhook and let your own stack take it from there.
Every ten minutes, every morning, or once a month — your competitor price sheet updates before you wake up.
Add one step to a page extraction and every row carries its title, description, image, author and canonical URL.
Name downloaded files after several columns at once — Brand, then Name, and they save as Brand_Product.
Browse a site as a tree and send the pages you pick straight to the extractor — no URL list to build first.
“Set it up in a couple of minutes and had a full product table before my coffee got cold.”
“The side panel picker is the fastest I've used — hover, click, and the table fills itself.”
“Connected it to my agent once — now price checks just happen every morning.”
Two ways in. Neither needs code.
No. You point at what you want on the page and click it — there are no selectors to write and nothing to install beyond the extension. Most people have their first export within minutes.
Not when you scrape locally — results sit in your browser until you export them. Only jobs you explicitly send to the cloud run on our infrastructure, in a workspace only you can see.
Your agent connects to your cloud workspace over MCP, where the scraping tools and your result tables live. You connect once by pasting a link; from then on you ask for data in plain English.
The extension installs free and local extraction is unlimited. Power features, exports, cloud runs and AI connections are part of the paid plans — see the pricing page for details.
Yes — it scrapes the page as your browser renders it, logged-in sessions included. Some heavily protected sites are better handled by the cloud's proxies.
Urban Web Scraper