> ## Documentation Index
> Fetch the complete documentation index at: https://docs.odella.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Extract Markdown Code Blocks

> Extract code blocks from markdown-formatted text

<Frame>
  <img src="https://mintcdn.com/odella/I7v1Fbrp2kx7rLoj/block-reference/assets/extract_markdown_code_blocks_dark.png?fit=max&auto=format&n=I7v1Fbrp2kx7rLoj&q=85&s=a80e75ba86408fc3549485e5da079726" alt="Extract Markdown Code Screenshot" className="block dark:hidden" width="888" height="386" data-path="block-reference/assets/extract_markdown_code_blocks_dark.png" />

  <img src="https://mintcdn.com/odella/I7v1Fbrp2kx7rLoj/block-reference/assets/extract_markdown_code_blocks_light.png?fit=max&auto=format&n=I7v1Fbrp2kx7rLoj&q=85&s=1108d89ae2ae6215a411ee075c9c23ee" alt="Extract Markdown Code Screenshot" className="hidden dark:block" width="878" height="400" data-path="block-reference/assets/extract_markdown_code_blocks_light.png" />
</Frame>

## Overview

The Extract Markdown Code Blocks block is designed to parse markdown-formatted text and extract all code blocks within it. This is particularly useful when working with documentation, README files, or any text that contains code snippets in markdown format.

## Inputs

<ParamField path="input" type="string" required>
  The markdown-formatted text from which to extract code blocks.
</ParamField>

## Outputs

<ResponseField name="firstBlock" type="string">
  The first code block found in the input markdown. If no code blocks are found, this output will be control-flow-excluded.
</ResponseField>

<ResponseField name="allBlocks" type="string[]">
  An array of strings, where each string represents a code block extracted from the input markdown.
</ResponseField>

<ResponseField name="languages" type="string[]">
  An array of strings containing the language identifiers specified for each code block. For example, in "\`\`\`python", "python" would be the language identifier.
</ResponseField>

## Example: Extracting Code from a README

1. Create a new flow and add an Extract Markdown Code Blocks block.
2. Add a Text block with markdown content containing code blocks and connect it to the `input` of the Extract Markdown Code Blocks block.
3. Connect any of the outputs to view the extracted code blocks or language information.
4. Run the flow to see the results.

## Error Handling

* If no code blocks are found in the input markdown, the `firstBlock` output will be control-flow-excluded and `allBlocks` will be an empty array.
* The block will not throw an error for invalid markdown; it will attempt to extract any code blocks it can identify.

## FAQ

<AccordionGroup>
  <Accordion title="What format are the code blocks expected to be in?">
    The block expects code blocks to be formatted using triple backticks (\`\`\`) as per standard markdown syntax. Both fenced code blocks with optional language identifiers are supported.
  </Accordion>

  <Accordion title="How are language identifiers handled?">
    Language identifiers are extracted from the opening fence of code blocks (e.g., \`\`\`python). If no language is specified, an empty string is used in the languages array.
  </Accordion>

  <Accordion title="Does this block handle inline code?">
    No, this block is designed to extract full code blocks, not inline code snippets that are typically wrapped in single backticks.
  </Accordion>
</AccordionGroup>

## See Also

* [Extract JSON Block](/block-reference/modifiers/extract-json)
* [Chunk Block](/block-reference/modifiers/chunk)
* [Text Block](/block-reference/data/text)