Extract fenced code blocks from Markdown or text
Paste Markdown or text into the input area. The tool finds fenced blocks that start with three or more backticks or tildes, then shows each block's order, language, 1-based source line range, and code. Parsing runs in browser JavaScript, and the extracted code is not executed.
How to extract and filter code blocks
- Paste text, select Upload and choose a
.md,.markdown, or.txtfile, or drop the file onto the input area. The shared text upload handler reads the first dropped file as text. - Review the result cards. The parser reruns when the input changes, and the stats area reports how many blocks were found.
- Use Language to choose All languages or a language from the input. Blocks without a language are shown as (no language) and can be filtered as (none). When filtered, the stats area also reports how many blocks are shown.
Copy or download extracted snippets
- Select Copy Shown to copy the visible blocks rebuilt with triple-backtick fences and the extracted first language token. It does not preserve the original fence character, fence length, or full info string.
- Select Download Shown to start one download for each visible block. Known language tags use mapped extensions; unknown tags are sanitized, and empty or unusable tags use
.txt. - Use the icon button titled Copy snippet to copy only one block's raw code, or the icon button titled Download snippet to download only that raw code.
Parser behavior and limitations
The parser uses a custom regular-expression pattern, not the full CommonMark algorithm. Opening fences use backticks or tildes with at least three characters. A closing fence must use the same character and at least the opening fence length, with optional surrounding whitespace. The first whitespace-delimited token after the opening fence is used as the language.
If a closing fence is missing, the block extends to the end of the input, is marked Unclosed fence, and uses the final input line as its end line. For non-empty input with no recognized fences, the results area says No fenced code blocks found. A language filter can also leave no visible cards when no block matches.
The local implementation has no explicit application-level input or file-size limit; browser constraints are not defined by this page.
Local processing behavior
The page-specific parsing code reads pasted text and selected files in browser JavaScript and does not upload them. That describes this parser's behavior, not an absolute site-wide privacy guarantee. For arbitrary delimiters, use Extract Text Between Delimiters.
Frequently Asked Questions (FAQs)
What counts as a fenced code block?
The parser looks for a line beginning with three or more backticks or tildes, optionally followed by an info string, and a later line using the same fence character with at least as many characters. It reads the first whitespace-delimited token as the language.
What happens to a block without a language?
It is displayed as (no language). In the Language selector, it is represented by (none).
What file types can I upload?
Upload accepts .md, .markdown, and .txt files. The selected file is read in the browser and its text is parsed.
What does Copy Shown copy?
It rebuilds the visible blocks with triple-backtick fences and the extracted first language token. It does not preserve the original fence character, fence length, or full info string.