Markdown tables in Jira: what converts and what does not

Updated 18 August 2026

Tables are where Markdown in Jira breaks down most visibly. A pipe table that renders perfectly in a README will usually arrive in Jira as a wall of vertical bars.

Why pipe tables are the hardest case

Autoformatting works line by line as you type. A Markdown table is not one line: it is a header row, a separator row, and body rows that only mean anything together. By the time the editor could tell it was a table, the rows are already plain paragraphs.

What to do instead

Build the table in Jira

Insert a table from the toolbar or the slash command, then fill the cells and use Tab to move between them. Tab at the end of the last row adds a row. This produces a real table node, which is what everything downstream expects.

Paste a rendered table

Render the Markdown first and copy the rendered table: from a preview pane, a README view, or a spreadsheet. Pasting rich table content usually produces a real Jira table. Pasting the pipe source never does.

Cell contents survive better than cell formatting. Check merged cells and column widths afterwards.

Data Center: use wiki markup

On Server and Data Center the table syntax is not pipes-with-dashes. Header cells use double bars and body cells single bars:

||Heading 1||Heading 2||
|cell A|cell B|
|cell C|cell D|

There is no separator row. If you paste Markdown pipe tables here they stay literal.

Things tables cannot do in Jira

• No nested tables.
• No per-column alignment from the Markdown separator row (:---: is ignored): alignment is a cell property set in the UI.
• Column widths are stored on the table, not expressed in text.

If tables are a daily problem

If you are moving tables from documents or spreadsheets into Jira often, a Markdown editor panel takes the pipe syntax directly and writes a real Jira table. Ours supports tables; see the box below.