will not work on newest articles because ar5iv is not indexed live (yet). At present, two weeks lag.

arxiv2markdown.py

# NB: ar5iv endpoint is lagged relative main arxiv
import subprocess


def convert_arxiv_url_to_markdown(arxiv_url, target_format='markdown_github'):
    parts = arxiv_url.split('/')
    article_id = parts[-1] if parts[-1] else parts[-2]  # Account for trailing slash
    article_id = article_id.replace('.pdf', '')
    html_url = f'https://ar5iv.labs.arxiv.org/html/{article_id}' 
    
    cmd = ['pandoc', '-t', target_format, '-o', '-', html_url]
    try:
        output = subprocess.run(cmd, check=True, capture_output=True, text=True)
        markdown_content = output.stdout
        return markdown_content
    except subprocess.CalledProcessError as e:
        return f"Error during conversion: {e}"


# demo
if __name__ == '__main__':
    arxiv_url = 'https://arxiv.org/abs/1910.06709'
    markdown = convert_arxiv_url_to_markdown(arxiv_url)
    print(markdown)