Markdown Table Extractor

Paste a Markdown document — the tool finds every GitHub-flavoured table, lets you pick one, and exports its data as CSV (RFC 4180), TSV, JSON (with type-coerced cells), or MySQL-flavour SQL (CREATE TABLE + INSERT). Each table is previewed live so you see exactly what will be exported before you copy. Pure client-side; nothing leaves your browser.

—

Detected tables

—

Paste Markdown above to detect tables.

Output format

Output format

Output

—
Text
 
 
 
Preview
 
 
 
 
  
 How parsing works & format trade-offs 

The detector reads the document line by line. A table starts
      with a header row of pipe-separated cells, followed
      immediately by a separator row of :---,
---:, :---:, or --- —
      one per column. Body rows continue until a blank line or a
      non-table line is hit. The header's most recent
# heading is captured as the table's name and is
      used by default for the SQL CREATE TABLE name.

 
 CSV quotes on need: a cell is wrapped in
"…" only when it contains the delimiter,
      a quote, or a line break. RFC 4180 quotes are doubled
      to escape them. TSV collapses every tab and
      newline inside a cell to a space. JSON
exports an array of objects keyed by the header; clean
      integers are emitted as numbers, true/false
as 1/0, otherwise the cell is left
      as a string. SQL infers each column's type
      from the body (INTEGER /
REAL / TEXT), quotes values with the
      appropriate literal, and emits one INSERT per
      50 rows. Markdown is a clean round-trip:
      padding is tuned to each column's widest cell and the
      alignment markers are preserved exactly.

 

Inline pipes inside a cell are not supported by GFM (the
      parser can't tell apart "a real column break" from "a literal
| in a cell"). Tables where this happens are
      flagged in the row count with a small warning so you know to
      edit the source. Cells inside backtick code spans
      (`a | b`) are honoured as the GFM spec requires.