diff --git a/migrations/20250219073939_cleanup_episodes.py b/migrations/20250219073939_cleanup_episodes.py
new file mode 100644
index 0000000..5a89730
--- /dev/null
+++ b/migrations/20250219073939_cleanup_episodes.py
@@ -0,0 +1,201 @@
+from datetime import datetime, timezone
+from typing import List, Optional, Tuple
+from dataclasses import dataclass
+
+from beanie import Document, free_fall_migration
+from pydantic import BaseModel, Field
+import PTT
+from tqdm.asyncio import tqdm
+import logging
+
+# Configure logging
+logging.basicConfig(
+    level=logging.INFO, format="%(asctime)s - %(levelname)s - [%(name)s] - %(message)s"
+)
+logger = logging.getLogger(__name__)
+
+
+class EpisodeFile(BaseModel):
+    season_number: int
+    episode_number: int
+    filename: Optional[str] = None
+
+
+class TorrentStreams(Document):
+    id: str
+    meta_id: str
+    torrent_name: str
+    size: int
+    episode_files: List[EpisodeFile] = Field(default_factory=list)
+    updated_at: datetime | None = Field(
+        default_factory=lambda: datetime.now(tz=timezone.utc)
+    )
+
+    class Settings:
+        name = "TorrentStreams"
+
+
+@dataclass
+class EpisodeDetail:
+    season_number: int
+    episode_number: int
+
+    def to_tuple(self) -> Tuple[int, int]:
+        return (self.season_number, self.episode_number)
+
+    def to_episode_file(self) -> EpisodeFile:
+        return EpisodeFile(
+            season_number=self.season_number, episode_number=self.episode_number
+        )
+
+
+class EpisodeProcessor:
+    def __init__(self, stream: TorrentStreams):
+        self.stream = stream
+        self.parsed_title = PTT.parse_title(stream.torrent_name)
+
+    def _parse_seasons_episodes(self) -> Tuple[List[int], List[int]]:
+        seasons = self.parsed_title.get("seasons", [])
+        episodes = self.parsed_title.get("episodes", [])
+
+        if not seasons and not episodes:
+            seasons = [1]
+
+        return seasons, episodes
+
+    def generate_episode_details(self) -> List[EpisodeDetail]:
+        seasons, episodes = self._parse_seasons_episodes()
+
+        if len(seasons) == 1 and episodes:
+            return [EpisodeDetail(seasons[0], episode) for episode in episodes]
+        elif len(seasons) == 1:
+            return [EpisodeDetail(seasons[0], 1)]
+        elif len(seasons) > 1:
+            return [EpisodeDetail(season, 1) for season in seasons]
+        elif episodes:
+            return [EpisodeDetail(1, episode) for episode in episodes]
+        else:
+            return [EpisodeDetail(1, 1)]
+
+    def process_episodes(self) -> List[EpisodeFile]:
+        existing_episodes = self.stream.episode_files
+        episode_details = self.generate_episode_details()
+
+        # Early return if no episode details generated
+        if not episode_details:
+            return existing_episodes
+
+        exist_any_filename = any(ep.filename for ep in existing_episodes)
+
+        if exist_any_filename:
+            return self._process_with_existing_filenames(
+                existing_episodes, episode_details
+            )
+        else:
+            return self._process_without_filenames(existing_episodes, episode_details)
+
+    def _process_with_existing_filenames(
+        self, existing_episodes: List[EpisodeFile], episode_details: List[EpisodeDetail]
+    ) -> List[EpisodeFile]:
+        cleaned_episodes = [ep for ep in existing_episodes if ep.filename]
+        cleaned_episodes_keys = {
+            (ep.season_number, ep.episode_number) for ep in cleaned_episodes
+        }
+
+        # Log removed episodes
+        removed_episodes = {
+            (ep.season_number, ep.episode_number)
+            for ep in existing_episodes
+            if not ep.filename
+        }
+        if removed_episodes:
+            logger.debug(
+                f"Removed empty filename episodes for stream {self.stream.id} "
+                f"with {removed_episodes}"
+            )
+
+        # Add missing episodes
+        new_episodes = [
+            detail.to_episode_file()
+            for detail in episode_details
+            if detail.to_tuple() not in cleaned_episodes_keys
+        ]
+        cleaned_episodes.extend(new_episodes)
+        return cleaned_episodes
+
+    def _process_without_filenames(
+        self, existing_episodes: List[EpisodeFile], episode_details: List[EpisodeDetail]
+    ) -> List[EpisodeFile]:
+        existing_episodes_keys = {
+            (ep.season_number, ep.episode_number) for ep in existing_episodes
+        }
+
+        new_episodes = [detail.to_episode_file() for detail in episode_details]
+
+        # Log changes
+        changes = [
+            detail
+            for detail in episode_details
+            if detail.to_tuple() not in existing_episodes_keys
+        ]
+        if changes:
+            logger.debug(f"Torrent {self.stream.id} Change episode details: {changes}")
+
+        return new_episodes
+
+
+class Forward:
+    @free_fall_migration(document_models=[TorrentStreams])
+    async def cleanup_episodes(self, session):
+        query = {
+            "episode_files": {
+                "$elemMatch": {
+                    "season_number": {"$exists": True},
+                    "$or": [{"filename": None}, {"filename": {"$exists": False}}],
+                }
+            },
+        }
+
+        streams = TorrentStreams.find(query)
+        count = await streams.count()
+        logger.info(f"Found {count} streams with empty filename episodes")
+
+        with tqdm(total=count, desc="Processing streams") as pbar:
+            async for stream in streams:
+                try:
+                    processor = EpisodeProcessor(stream)
+                    cleaned_episodes = processor.process_episodes()
+
+                    # Skip update if no changes
+                    if cleaned_episodes == stream.episode_files:
+                        pbar.update(1)
+                        continue
+
+                    stream.episode_files = cleaned_episodes
+                    new_keys = [
+                        (ep.season_number, ep.episode_number)
+                        for ep in stream.episode_files
+                    ]
+
+                    logger.debug(
+                        f"Updated episodes for stream {stream.id} '{stream.torrent_name}' "
+                        f"with {new_keys}"
+                    )
+
+                    stream.updated_at = datetime.now(tz=timezone.utc)
+                    await stream.save()
+
+                except Exception as e:
+                    logger.error(
+                        f"Error processing stream {stream.id}: {str(e)}", exc_info=True
+                    )
+                    continue
+
+                pbar.update(1)
+
+        logger.info("Migration completed successfully")
+
+
+class Backward:
+
+    pass