yt-dlp/yt-dlp · error · ExtractorError

Unable to extract embedUrl

Error message

Unable to extract embedUrl

What it means

Thrown by NDRIE._real_extract (yt_dlp/extractor/ndr.py:136). NDR articles expose media via a direct embed URL, a 'sophora ID' short code (pattern ^[a-z]+\d+$) that triggers a secondary page-path/image scan, or the NDR info-player template; when embed_url is still falsy after all of those fallbacks, extraction stops.

Source

Thrown at yt_dlp/extractor/ndr.py:136

                default=None)
            or self._search_regex(
                r'\bembedUrl["\']\s*:\s*(["\'])(?P<url>(?:(?!\1).)+)\1', webpage,
                'embed URL', group='url', default=None)
            or self._search_regex(
                r'\bvar\s*sophoraID\s*=\s*(["\'])(?P<url>(?:(?!\1).)+)\1', webpage,
                'embed URL', group='url', default=''))
        # some more work needed if we only found sophoraID
        if re.match(r'^[a-z]+\d+$', embed_url):
            # get the initial part of the url path,. eg /panorama/archiv/2022/
            parsed_url = urllib.parse.urlparse(url)
            path = self._search_regex(rf'(.+/){display_id}', parsed_url.path or '', 'embed URL', default='')
            # find tell-tale image with the actual ID
            ndr_id = self._search_regex(rf'{path}([a-z]+\d+)(?!\.)\b', webpage, 'embed URL', default=None)
            # or try to use special knowledge!
            NDR_INFO_URL_TPL = 'https://www.ndr.de/info/%s-player.html'
            embed_url = f'ndr:{ndr_id}' if ndr_id else NDR_INFO_URL_TPL % (embed_url, )
        if not embed_url:
            raise ExtractorError('Unable to extract embedUrl')

        description = self._search_regex(
            r'<p[^>]+itemprop="description">([^<]+)</p>',
            webpage, 'description', default=None) or self._og_search_description(webpage)
        timestamp = parse_iso8601(
            self._search_regex(
                (r'<span[^>]+itemprop="(?:datePublished|uploadDate)"[^>]+content="(?P<cont>[^"]+)"',
                 r'\bvar\s*pdt\s*=\s*(?P<q>["\'])(?P<cont>(?:(?!(?P=q)).)+)(?P=q)'),
                webpage, 'upload date', group='cont', default=None))
        info = self._search_json_ld(webpage, display_id, default={})
        return merge_dicts({
            '_type': 'url_transparent',
            'url': embed_url,
            'display_id': display_id,
            'description': description,
            'timestamp': timestamp,
        }, info)

View on GitHub (pinned to 81ecd58b13)

Solutions

1. Confirm in a browser that the article actually contains playable video or audio
2. Update yt-dlp to latest nightly - NDR markup changes are common and patched
3. Try the canonical /mediathek/ or ndr: <doc-id> form of the same content
4. Open devtools, find the real iframe/video src on the page, and pass that URL directly
5. Report with --verbose if the page has media but extraction still fails

import re, urllib.request

def ndr_page_has_media(url) -> bool:
    with urllib.request.urlopen(urllib.request.Request(url, headers={'User-Agent': UA})) as r:
        html = r.read().decode('utf-8', 'replace')
    return bool(re.search(r'(data-url=|embed|mediathek|<iframe)', html))

from yt_dlp.utils import ExtractorError
try:
    info = ydl.extract_info(url, download=False)
except ExtractorError as e:
    if 'Unable to extract embedUrl' in str(e):
        # all fallbacks failed: likely text-only article or markup change
        skip_article(url)

When it happens

Trigger: An ndr.de/sportschau article page where the embed-URL regex, the path+image-based sophora-ID lookup, and the info-player fallback all fail: text-only article, JS-rendered player, or a template change that moved the media references.

Common situations: Article links that contain no A/V media at all; NDR redesigns (happened repeatedly - the layered fallbacks in this function are scar tissue); audio items handled by a sibling extractor; JS-only pages served to bots.

Related errors

• No video metadata found in webpage
• No playlist or embed found in webpage
• Unsafe placeholder for exec command: {na!r} The --output-na-
• Unsafe default(s) in exec command: {outtmpl!r} Conversions a
• Invalid syntax in Cookie Header

AI-assisted analysis of yt-dlp/yt-dlp@81ecd58b13 (2026-08-22). Data as JSON: /api/errors/2614aa27c1695d9e. Report an issue: GitHub.