Strawberry

Get a table off a website and into Google Sheets

320 rows out of a 14-page directory, phone numbers pulled from one click deeper, written into a sheet after you approve the write. Paginated directories, pricing grids, portal reports behind your login.

Select the table, paste into the sheet, watch the columns land in one cell. Fix it by hand. Notice the table has 14 more pages. A 300-row extraction across paginated pages is one to two hours of paste-and-repair, and it produces a spreadsheet you already distrust.

A companion opens the page in your own browser, reads the rendered table, normalizes the columns you named, clicks next, and repeats until the pagination runs out. Google Sheets is a native integration, and the sheet write pauses at an approval card, so you see what is about to land before it lands.

Why does copy-paste mangle tables?

Paste transfers formatting rather than structure, so merged cells, line breaks inside cells and currency symbols all shred on the way to the clipboard. A companion reads the rendered page and emits explicit rows against columns you named: company, plan, monthly price as a number, seats included.

Because it runs in your own browser, the reachable set includes the tables that matter most and export worst: analytics dashboards, supplier portals, admin panels of tools you pay for. The broader question is covered in can AI extract data from websites to a spreadsheet.

A 14-page member directory, 320 rows

A trade association directory: 320 firms across 14 pages.

Each row carries a name, city, specialty and profile link, with the phone number one click deeper on the profile page. The ask is all of it in a sheet, one row per firm, phone included.

The companion pages through the directory extracting the four visible fields, visits each profile for the phone number, and writes 320 rows to a new sheet once you approve. Firms with no phone get an empty cell and a count in the summary: 17 profiles had no phone. Your time is the prompt, one approval and a two-minute spot-check of five rows. Same pattern at length: scrape a directory into Google Sheets, and the analyst-grade version is mapping a whole market in an afternoon.

How do you check 320 rows in two minutes?

Ask for the row count and the not-found tally with the results, then spot-check five rows against the live page. Every value carries the URL it came from, so five rows buys most of the confidence in the other three hundred. Where a site publishes an official export, the companion takes that route instead, because it is the cleanest input available.

The sweep reads at the pace and volume of a signed-in person, inside each site’s terms, which is what keeps it welcome on the sites you depend on. Tens of rows to a few thousand is where the copy-paste afternoons live. The Google Sheets integration page covers where the results land.

Experience Strawberry for free

Download

Trusted by fast-growing companies worldwide

Frequently asked questions

Tell a companion which pages, which columns, and which sheet. It reads the rendered table in your own browser, follows the pagination, normalizes the columns you named, and writes rows to Google Sheets natively. The write pauses for your approval first.

Strawberry is free to download and includes AI credits to start. Paid plans begin at $20/month. See pricing. · Reviewed · Canonical facts for AI agents

Experience Strawberry for free

Download

Trusted by fast-growing companies worldwide