How to paste a ChatGPT table into Excel or Google Sheets

What the chat window shows you is a rendered table. What your clipboard receives is usually this:

| Region | Q1 | Q2 |
| --- | ---: | ---: |
| Europe | 412 | 486 |
| Asia | 377 | 381 |

That is a Markdown pipe table. The vertical bars are ordinary characters that look like column dividers to a human reader. The second line, with its dashes and colons, tells a Markdown renderer where the header ends and how to align each column.

Excel splits pasted text on tab characters, not pipes. Since there are no tabs, each line is a single value, and the entire table stacks into column A with the pipes still visible. Google Sheets and Numbers behave identically.

The alignment row is the giveaway that this was never spreadsheet data. No spreadsheet export contains a row of hyphens and colons.

Rather than pasting and repairing, convert the table into a format spreadsheets already understand.

For Excel and Numbers, Markdown Table to Excel produces a genuine .xlsx file. Paste the table, check the cell preview, download the spreadsheet, open it. Numbers stay numbers, the header row stays a header row, and nothing needs splitting afterwards. You can paste the whole answer rather than hunting for the table: every pipe table in it becomes its own sheet. And if you would rather not download anything, Copy puts tab-separated cells on the clipboard, which Excel, Sheets and Numbers all paste as a grid.

For Google Sheets, Markdown Table to CSV is usually smoother. CSV imports natively through File → Import, and you avoid the download-then-upload round trip that XLSX requires in a browser. The converter escapes quotes and embedded commas properly, which is where hand-rolled conversions typically break.

Both run locally in your browser, so the data never leaves your machine — worth noting if the table contains anything you would not paste into a random website.

If you would rather not leave the spreadsheet, two built-in routes work.

Text to Columns (Excel). Paste the raw table into column A, select it, then Data → Text to Columns → Delimited, and set the delimiter to the pipe character. You will still need to delete the alignment row and trim the stray empty columns created by the leading and trailing pipes.

SPLIT (Google Sheets). With the raw table in column A, use =SPLIT(A1, "|") and fill down, then copy and paste-as-values. Same cleanup applies.

These are fine for a single small table. They get tedious quickly, and both leave whitespace padding inside every cell — the spaces Markdown uses for readability become part of the value, so Europe will not match Europe in a lookup formula. That trailing-space problem is the most common reason a converted table looks right but behaves wrong.

A note on copying: if you drag-select the rendered table in the chat window rather than using the copy button, some browsers put tab-separated text on the clipboard, which pastes into columns correctly. It is inconsistent across browsers and chat interfaces, so it is worth trying but not worth relying on.

Sheets fails the same way Excel does — pipes are not a delimiter it recognises, so every row lands in one cell of column A — but the routes out of it are slightly different, and one of them is genuinely convenient.

Split text to columns. Paste the raw table into a column, keep the pasted range selected, and use Data → Split text to columns. Sheets offers a separator choice, including a custom one, so you can set it to the pipe character. It also shows a small control near the pasted data straight after any paste of delimited text, which does the same thing without the menu trip. As with Excel's Text to Columns, you are left with the alignment row to delete, empty columns at each edge from the leading and trailing pipes, and a space of padding inside every value.

Import a CSV instead. File → Import parses a file rather than a paste, and lets you say what the separator is and whether to convert text to numbers and dates. That is the difference that matters: an import decides the column types deliberately, where a paste guesses. Run the table through Markdown Table to CSV first and there is nothing left to guess about — the alignment row is gone, padding is trimmed, and commas inside cells are quoted correctly.

Sheets as a staging post. Once a table is in real cells, copying that range carries the grid with it. Pasting from Sheets into Excel, Word, Slides, Docs, or an email keeps the rows and columns intact, because the clipboard now holds tabular data rather than a line of text with bars in it. If a table has to reach an app with no import of its own, going through a spreadsheet first is more reliable than fighting the destination.

One Sheets-specific thing to watch: values like 3-4 or 1/2 in a converted cell will be read as dates. If a column arrives full of dates you did not ask for, undo, format the destination column as plain text, and paste again.

Often the table you want is one part of a much longer response, mixed in with explanation above and below it.

Extract Tables from Text finds table-like rows inside surrounding prose — pipe-delimited or tab-delimited — and returns just the table as CSV. Paste the entire answer and take the data out, rather than hand-selecting rows and hoping you caught the boundaries.

Two things worth checking on any converted table before you build on it:

• Numbers stored as text. If a column arrives left-aligned in Excel, the values are text. Currency symbols, thousands separators, and percent signs in the original are the usual cause. Select the column and use Data → Text to Columns → Finish to coerce them.
• Merged or wrapped cells. Markdown tables cannot express merged cells, so a model presenting complex data will sometimes fake it with repeated values or line breaks inside a cell. Those need a human decision, not a converter.

If the rest of the response also needs to travel — headings, lists, the explanation around the table — see why pasted AI text shows ** and ##.