diff --git a/backend/download/src/yt_dlp_base.py b/backend/download/src/yt_dlp_base.py
index 909931c0..fa3c2b03 100644
--- a/backend/download/src/yt_dlp_base.py
+++ b/backend/download/src/yt_dlp_base.py
@@ -4,6 +4,7 @@ functionality:
 - handle yt-dlp errors
 """
 
+import re
 from datetime import datetime
 from http import cookiejar
 from io import StringIO
@@ -14,6 +15,114 @@ from common.src.ta_redis import RedisArchivist
 from django.conf import settings
 
 
+class ExtractorArgsParser:
+    """
+    Parse extractor_args string following yt-dlp format.
+
+    Format: "extractor1:key1=val1,val2;key2=val3 extractor2:key3=val4"
+
+    - Whitespace separates multiple extractors
+    - Colon (:) separates extractor name from key-value pairs
+    - Semicolon (;) separates multiple key-value pairs
+    - Equals (=) separates key from values
+    - Comma (,) separates multiple values (escaped commas \\, are preserved)
+
+    Example:
+        >>> parser = ExtractorArgsParser()
+        >>> result = parser.parse("youtube:key1=val1,val2;key2=val3")
+        >>> # Returns: {"youtube": {"key1": ["val1", "val2"], "key2": ["val3"]}}
+    """
+
+    @staticmethod
+    def _parse_key_values(key, vals=''):
+        """
+        Parse a single key=values pair following yt-dlp logic.
+
+        Args:
+            key: The key name (will be normalized)
+            vals: Comma-separated values (escaped commas preserved)
+
+        Returns:
+            tuple: (normalized_key, list_of_values)
+        """
+        # Normalize key: strip, lowercase, replace hyphens with underscores
+        normalized_key = key.strip().lower().replace('-', '_')
+
+        # Split values by comma, but preserve escaped commas
+        # yt-dlp uses: re.split(r'(?<!\\),', vals)
+        if not vals:
+            return normalized_key, []
+
+        value_list = [
+            val.replace(r'\,', ',').strip()
+            for val in re.split(r'(?<!\\),', vals)
+        ]
+
+        return normalized_key, value_list
+
+    def parse(self, extractor_args_str):
+        """
+        Parse extractor_args string into nested dictionary structure.
+
+        Args:
+            extractor_args_str: String in format "extractor:key=val;key2=val2"
+
+        Returns:
+            dict: Nested dictionary with structure:
+                  {extractor_name: {key: [values]}}
+
+        Example:
+            >>> parser.parse("youtube:lang=en,es;key2=val generic:key3=val3")
+            {'youtube': {'lang': ['en', 'es'], 'key2': ['val']},
+             'generic': {'key3': ['val3']}}
+        """
+        if not extractor_args_str or not extractor_args_str.strip():
+            return {}
+
+        result = {}
+
+        # Split by whitespace to get individual extractor specifications
+        extractor_specs = extractor_args_str.strip().split()
+
+        for spec in extractor_specs:
+            # Split by first colon to separate extractor from args
+            if ':' not in spec:
+                print(f"[extractor_args] Warning: Invalid format (missing colon): {spec}")
+                continue
+
+            extractor_name, args_part = spec.split(':', 1)
+            extractor_name = extractor_name.strip()
+
+            if not extractor_name:
+                print(f"[extractor_args] Warning: Empty extractor name in: {spec}")
+                continue
+
+            # Initialize extractor dict if not exists
+            if extractor_name not in result:
+                result[extractor_name] = {}
+
+            # Split args by semicolon to get individual key=value pairs
+            key_value_pairs = args_part.split(';')
+
+            for pair in key_value_pairs:
+                if not pair.strip():
+                    continue
+
+                # Split by first equals sign
+                if '=' not in pair:
+                    print(f"[extractor_args] Warning: Invalid pair (missing =): {pair}")
+                    continue
+
+                key_part, value_part = pair.split('=', 1)
+                normalized_key, value_list = self._parse_key_values(
+                    key_part, value_part
+                )
+
+                result[extractor_name][normalized_key] = value_list
+
+        return result
+
+
 class YtWrap:
     """wrap calls to yt"""
 
@@ -37,6 +146,7 @@ class YtWrap:
         if self.config:
             self._add_cookie()
             self._add_potoken()
+            self._add_extractor_args()
 
         if getattr(settings, "DEBUG", False):
             del self.obs["quiet"]
@@ -63,6 +173,32 @@ class YtWrap:
                 }
             )
 
+    def _add_extractor_args(self):
+        """add custom extractor_args from config if present"""
+        extractor_args_str = self.config["downloads"].get("extractor_args")
+        if not extractor_args_str:
+            return
+
+        # Parse the extractor_args string
+        parser = ExtractorArgsParser()
+        parsed_args = parser.parse(extractor_args_str)
+
+        if not parsed_args:
+            return
+
+        # Merge with existing extractor_args if present
+        if "extractor_args" not in self.obs:
+            self.obs["extractor_args"] = {}
+
+        for extractor_name, args_dict in parsed_args.items():
+            if extractor_name not in self.obs["extractor_args"]:
+                self.obs["extractor_args"][extractor_name] = {}
+
+            # Merge the arguments, with parsed args taking precedence
+            self.obs["extractor_args"][extractor_name].update(args_dict)
+
+        print(f"[extractor_args] Applied custom args: {parsed_args}")
+
     def download(self, url):
         """make download request"""
         self.obs.update({"check_formats": "selected"})