Table Grabber — Copy, Edit & Export Any Table to CSV, Excel

Table Grabber copies any table on any web page into the format you actually need. Every format is free. Every row is included. There is no account, no row limit, no watermark, no trial, and no upload — the extension contains no network code at all. Click the toolbar icon and it lists every table it can find on the page, with the row and column count of each. Copy one straight away, or open it in the editor to clean it up first. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ THREE WAYS TO GRAB SOMETHING CLICK IT Point at the table you want and click. The selection is outlined as you move, with its row and column count, and the arrow keys widen or narrow it if the outline lands on the wrong container. Any element can be selected — if it is not a table, it is converted into one. DRAG A BOX ROUND IT Not everything that looks like a table is one. A price list built with flexbox, a block of stat tiles, a receipt in fixed-width text: none of these have rows or columns in the markup, only positions on the screen — which is exactly what you are reading them by. Drag a box and the text inside it is rebuilt into a table from those positions, coping with right-aligned number columns, cells of different heights, and gaps where a value is simply missing. If the box covers only part of a real table, both readings are offered with their row and column counts: just the part you covered, or the whole table. Neither is guessed at silently. PASTE IT The editor takes data you bring yourself: delimited text, structured data, a table written in Markdown, or table markup copied from anywhere. Drop a file onto it and it reads that too. Every grab ends the same way — a panel showing what was found and how it was read, with copy buttons and a way back, including the attempts that find nothing. Nothing leaves you staring at a dimmed page wondering whether it worked. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ FINDS THE TABLES OTHER TOOLS MISS • Ordinary HTML tables, including ones nested inside other tables • Tables inside open shadow roots and same-origin iframes, which a plain search of the page never sees • Grids built out of plain boxes — the kind modern web apps draw with CSS and JavaScript, which contain no table markup whatsoever. These are found by their repeating structure rather than by hardcoding any particular grid library's class names, so it works on hand-rolled grids and the popular libraries alike • Virtualised grids that keep only the visible rows loaded and swap them as you scroll. Table Grabber scrolls the grid in steps and accumulates the rest, placing each row by the grid's own row numbering where it publishes one • Paginated tables: it finds the Next control, clicks it, waits for the table to actually change rather than guessing at a delay, and merges every page into a single export. It stops at the last page, and you can stop it sooner Cross-origin frames are skipped. The browser correctly forbids reading them, and no extension should be trying. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ GETS THE SHAPE RIGHT MERGED CELLS This is the reason most table extractors quietly produce misaligned data. When one cell spans two rows, a reader that walks the cells in order shifts every column after it — so your numbers end up filed under the wrong headings, and nothing warns you. You find out later, in the spreadsheet, if you find out at all. Table Grabber lays every cell onto a coordinate grid before reading anything, so a merged cell still occupies the positions it visually occupies. You choose what happens in those positions: repeat the value down, leave them blank, or mark them explicitly. MULTI-ROW HEADERS A grouped header — one "Region" cell sitting above "Country" and "City" — is two header rows. Take only the first and you lose the columns; take only the second and you lose the group. Header rows are detected and stacked into one name per column, so you get "Region – Country" instead of half the information. NUMBERS THAT STAY NUMBERS $1,234.00 pasted into a spreadsheet is text, and SUM() over a column of text returns zero. Table Grabber detects each column's type — number, currency, percent, date, boolean, URL, text — and can rewrite the values so the destination reads them as numbers. It understands both 1,234.56 and the European 1.234,56. It handles accounting negatives written as (1,234). And it survives placeholder values: a column of prices with a few N/A entries is still recognised as numeric, rather than being written off as text because of them. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ OUTPUT FORMATS Thirteen of them, covering spreadsheets, structured data, databases and documents. The screenshots below show the full list, and the guide linked at the end describes each one. Each one is written properly rather than approximated. The spreadsheet export is a real workbook with numeric cells, a frozen header row and columns sized to their contents — so formulas and charts work the moment it opens, with no import wizard and no re-typing. Delimited text is quoted to the standard and carries a byte order mark, so accented characters arrive intact instead of as mojibake. The structured formats keep genuine numbers and booleans instead of quoting everything, and turn the headers into usable keys. The database export infers a column type per field and batches its inserts. The document formats come out aligned in the source, with the special characters escaped. PASTE THAT LANDS IN CELLS Copying to the clipboard puts table markup there alongside the plain text. Google Sheets and Excel pick that up and give you real cells, instead of dumping the whole table into A1 as one long string. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ EDIT AND REARRANGE IT FIRST Most table extractors hand you the data exactly as the page had it, and any fix — a typo at the source, a row in the wrong place, one line of junk in the middle — sends you to a spreadsheet to do by hand. This one lets you finish the job before the …