diff --git a/CHANGELOG.md b/CHANGELOG.md
index 2df979fe..78e2c3f7 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -13,6 +13,7 @@ and this project adheres to
 - 🚀(front) fix frontend build failure
 - 🐛(makefile) fix args in make test
 - 🩹(backend) fix case-insensitive email deduplication in merge command
+- 🐛(summary) support media files with bad streams #1478
 
 ## [1.22.0] - 2026-07-03
 
diff --git a/src/summary/summary/core/file_service.py b/src/summary/summary/core/file_service.py
index 9088ccc9..1b4437f4 100644
--- a/src/summary/summary/core/file_service.py
+++ b/src/summary/summary/core/file_service.py
@@ -142,6 +142,7 @@ class MediaInfo:
     has_video: bool
     audio_duration_seconds: float | None
     audio_codec_name: str | None
+    has_bad_stream: bool = False
 
 
 def get_media_info(local_path: Path) -> MediaInfo:
@@ -174,8 +175,9 @@ def get_media_info(local_path: Path) -> MediaInfo:
     data = json.loads(result.stdout)
 
     streams = data.get("streams", [])
-    has_audio = any(el["codec_type"] == "audio" for el in streams)
-    has_video = any(el["codec_type"] == "video" for el in streams)
+    has_audio = any(el.get("codec_type") == "audio" for el in streams)
+    has_video = any(el.get("codec_type") == "video" for el in streams)
+    has_bad_stream = any(el.get("codec_type", None) is None for el in streams)
     audio_codec_name = next(
         (
             stream.get("codec_name")
@@ -193,10 +195,11 @@ def get_media_info(local_path: Path) -> MediaInfo:
         has_video=has_video,
         audio_duration_seconds=audio_duration_seconds,
         audio_codec_name=audio_codec_name,
+        has_bad_stream=has_bad_stream,
     )
 
 
-def extract_audio_from_video(media_info: MediaInfo) -> Path:
+def extract_audio_from_media(media_info: MediaInfo) -> Path:
     """Extracts the audio track from a video file and saves it as a separate audio file.
 
     Based on the provided audio codec,
@@ -451,7 +454,12 @@ class FileService:
 
             if media_info.has_video:
                 logger.info("Video file detected, extracting audio...")
-                processed_path = extract_audio_from_video(media_info)
+                processed_path = extract_audio_from_media(media_info)
+            # Bad streams may cause transcription issues on WhisperX,
+            # So we extract the audio properly
+            elif media_info.has_bad_stream:
+                logger.info("Bad stream detected, extracting audio...")
+                processed_path = extract_audio_from_media(media_info)
             else:
                 processed_path = downloaded_path
 
diff --git a/src/summary/tests/unit/test_file_service.py b/src/summary/tests/unit/test_file_service.py
index 8f9a92dd..f7e6fc6d 100644
--- a/src/summary/tests/unit/test_file_service.py
+++ b/src/summary/tests/unit/test_file_service.py
@@ -1,12 +1,15 @@
 """Unit tests for the file service."""
 
+import json
 from pathlib import Path
+from unittest.mock import Mock
 
 import pytest
 
+from summary.core import file_service
 from summary.core.file_service import (
     MediaInfo,
-    extract_audio_from_video,
+    extract_audio_from_media,
     get_media_info,
 )
 
@@ -110,12 +113,40 @@ def test_media_info_invalid_file() -> None:
     )
 
 
+def test_media_info_ignores_empty_stream_entry(monkeypatch: pytest.MonkeyPatch) -> None:
+    """Test stream parsing when ffprobe returns an empty stream object."""
+    ffprobe_payload = {
+        "programs": [],
+        "stream_groups": [],
+        "streams": [
+            {"codec_name": "vorbis", "codec_type": "audio"},
+            {},
+        ],
+    }
+
+    run_mock = Mock(
+        return_value=Mock(stdout=json.dumps(ffprobe_payload), stderr="", returncode=0)
+    )
+    monkeypatch.setattr(file_service.subprocess, "run", run_mock)
+    monkeypatch.setattr(
+        file_service, "get_media_duration_seconds", Mock(return_value=2.5)
+    )
+
+    media_info = get_media_info(BASE_PATH / "audio-sample-android-firefox.ogg")
+
+    assert media_info.has_audio is True
+    assert media_info.has_video is False
+    assert media_info.has_bad_stream is True
+    assert media_info.audio_codec_name == "vorbis"
+    assert media_info.audio_duration_seconds == 2.5
+
+
 def test_extract_audio_from_video():
     """Test that extract_audio_from_video can extract audio from a video file."""
     path = None
     # A bit of cleanup logic since this is not a generator
     try:
-        path = extract_audio_from_video(MEDIA_INFO_SAMPLE_VISIO)
+        path = extract_audio_from_media(MEDIA_INFO_SAMPLE_VISIO)
         assert path.name.endswith(".m4a")
     except Exception as e:
         pytest.fail(f"Failed to extract audio from video: {e}")