diff --git a/yt_dlp/extractor/extractors.py b/yt_dlp/extractor/extractors.py
index 9dd7543cf3..1d2f2a1b40 100644
--- a/yt_dlp/extractor/extractors.py
+++ b/yt_dlp/extractor/extractors.py
@@ -1126,6 +1126,10 @@ from .pinterest import (
     PinterestIE,
     PinterestCollectionIE,
 )
+from .pixivsketch import (
+    PixivSketchIE,
+    PixivSketchUserIE,
+)
 from .pladform import PladformIE
 from .planetmarathi import PlanetMarathiIE
 from .platzi import (
diff --git a/yt_dlp/extractor/pixivsketch.py b/yt_dlp/extractor/pixivsketch.py
new file mode 100644
index 0000000000..f0ad0b24a1
--- /dev/null
+++ b/yt_dlp/extractor/pixivsketch.py
@@ -0,0 +1,122 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+from .common import InfoExtractor
+from ..utils import (
+    ExtractorError,
+    traverse_obj,
+    unified_timestamp,
+)
+
+
+class PixivSketchBaseIE(InfoExtractor):
+    def _call_api(self, video_id, path, referer, note='Downloading JSON metadata'):
+        response = self._download_json(f'https://sketch.pixiv.net/api/{path}', video_id, note=note, headers={
+            'Referer': referer,
+            'X-Requested-With': referer,
+        })
+        errors = traverse_obj(response, ('errors', ..., 'message'))
+        if errors:
+            raise ExtractorError(' '.join(f'{e}.' for e in errors))
+        return response.get('data') or {}
+
+
+class PixivSketchIE(PixivSketchBaseIE):
+    IE_NAME = 'pixiv:sketch'
+    _VALID_URL = r'https?://sketch\.pixiv\.net/@(?P<uploader_id>[a-zA-Z0-9_-]+)/lives/(?P<id>\d+)/?'
+    _TESTS = [{
+        'url': 'https://sketch.pixiv.net/@nuhutya/lives/3654620468641830507',
+        'info_dict': {
+            'id': '7370666691623196569',
+            'title': 'まにあえクリスマス！',
+            'uploader': 'ぬふちゃ',
+            'uploader_id': 'nuhutya',
+            'channel_id': '9844815',
+            'age_limit': 0,
+            'timestamp': 1640351536,
+        },
+        'skip': True,
+    }, {
+        # these two (age_limit > 0) requires you to login on website, but it's actually not required for download
+        'url': 'https://sketch.pixiv.net/@namahyou/lives/4393103321546851377',
+        'info_dict': {
+            'id': '4907995960957946943',
+            'title': 'クリスマスなんて知らん🖕',
+            'uploader': 'すゃもり',
+            'uploader_id': 'suya2mori2',
+            'channel_id': '31169300',
+            'age_limit': 15,
+            'timestamp': 1640347640,
+        },
+        'skip': True,
+    }, {
+        'url': 'https://sketch.pixiv.net/@8aki/lives/3553803162487249670',
+        'info_dict': {
+            'id': '1593420639479156945',
+            'title': 'おまけ本作業(リョナ有)',
+            'uploader': 'おぶい / Obui',
+            'uploader_id': 'oving',
+            'channel_id': '17606',
+            'age_limit': 18,
+            'timestamp': 1640330263,
+        },
+        'skip': True,
+    }]
+
+    def _real_extract(self, url):
+        video_id, uploader_id = self._match_valid_url(url).group('id', 'uploader_id')
+        data = self._call_api(video_id, f'lives/{video_id}.json', url)
+
+        if not traverse_obj(data, 'is_broadcasting'):
+            raise ExtractorError(f'This live is offline. Use https://sketch.pixiv.net/@{uploader_id} for ongoing live.', expected=True)
+
+        m3u8_url = traverse_obj(data, ('owner', 'hls_movie', 'url'))
+        formats = self._extract_m3u8_formats(
+            m3u8_url, video_id, ext='mp4',
+            entry_protocol='m3u8_native', m3u8_id='hls')
+        self._sort_formats(formats)
+
+        return {
+            'id': video_id,
+            'title': data.get('name'),
+            'formats': formats,
+            'uploader': traverse_obj(data, ('user', 'name'), ('owner', 'user', 'name')),
+            'uploader_id': traverse_obj(data, ('user', 'unique_name'), ('owner', 'user', 'unique_name')),
+            'channel_id': str(traverse_obj(data, ('user', 'pixiv_user_id'), ('owner', 'user', 'pixiv_user_id'))),
+            'age_limit': 18 if data.get('is_r18') else 15 if data.get('is_r15') else 0,
+            'timestamp': unified_timestamp(data.get('created_at')),
+            'is_live': True
+        }
+
+
+class PixivSketchUserIE(PixivSketchBaseIE):
+    IE_NAME = 'pixiv:sketch:user'
+    _VALID_URL = r'https?://sketch\.pixiv\.net/@(?P<id>[a-zA-Z0-9_-]+)/?'
+    _TESTS = [{
+        'url': 'https://sketch.pixiv.net/@nuhutya',
+        'only_matching': True,
+    }, {
+        'url': 'https://sketch.pixiv.net/@namahyou',
+        'only_matching': True,
+    }, {
+        'url': 'https://sketch.pixiv.net/@8aki',
+        'only_matching': True,
+    }]
+
+    @classmethod
+    def suitable(cls, url):
+        return super(PixivSketchUserIE, cls).suitable(url) and not PixivSketchIE.suitable(url)
+
+    def _real_extract(self, url):
+        user_id = self._match_id(url)
+        data = self._call_api(user_id, f'lives/users/@{user_id}.json', url)
+
+        if not traverse_obj(data, 'is_broadcasting'):
+            try:
+                self._call_api(user_id, 'users/current.json', url, 'Investigating reason for request failure')
+            except ExtractorError as ex:
+                if ex.cause and ex.cause.code == 401:
+                    self.raise_login_required(f'Please log in, or use direct link like https://sketch.pixiv.net/@{user_id}/1234567890', method='cookies')
+            raise ExtractorError('This user is offline', expected=True)
+
+        return self.url_result(f'https://sketch.pixiv.net/@{user_id}/lives/{data["id"]}')