Example Use Cases

We created example use cases for you. You can check and get inspired by visiting our GitHub repository.

Searching for Keywords in Channel Transcripts

One common use case is searching for specific keywords or phrases across all videos in a channel. Here's an example:

from ytfetcher import YTFetcher
from ytfetcher.services import JSONExporter

def search_in_channel(handle: str, keyword: str):
    # 1. Initialize using your factory method
    fetcher = YTFetcher.from_channel(
        channel_handle=handle,
        max_results=5
    )

    # 2. Fetch the structured data
    print(f"Searching for '{keyword}' in @{handle}...")
    channel_data = fetcher.fetch_youtube_data()

    matches = []

    # 3. Iterate through your ChannelData models
    for video in channel_data:
        for entry in video.transcripts:
            if keyword.lower() in entry.text.lower():
                matches.append({
                    "video_id": video.video_id,
                    "timestamp": entry.start,
                    "text": entry.text
                })

    # 4. Show off the results
    print(f"Found {len(matches)} matches!")
    for m in matches[:5]: # Show first 5
        print(f"[{m['timestamp']}s] in {m['video_id']}: {m['text']}")

    # 5. Use your JSONExporter to save the findings
    exporter = JSONExporter(
        channel_data=channel_data,
        filename=f"{keyword}_search_results",
        timing=True
    )
    exporter.write()

if __name__ == "__main__":
    search_in_channel("TheOffice", "Dunder Mifflin")

Sentiment Analysis With Youtube Comments

We can use fetch_with_comments method to get top comments and use it for sentiment analysis.

First install vaderSentiment for analysis.

pip install vaderSentiment

Here's the full code:

from ytfetcher import YTFetcher
from vaderSentiment.vaderSentiment import SentimentIntensityAnalyzer

def analyze_comments(handle: str):
    # 1. Initialize the fetcher for a specific channel
    fetcher = YTFetcher.from_channel(channel_handle=handle, max_results=3)

    # 2. Fetch data WITH comments (using your specific method)
    print(f"Fetching comments from @{handle}...")
    channel_data = fetcher.fetch_with_comments(max_comments=20)

    analyzer = SentimentIntensityAnalyzer()
    stats = {"Positive": 0, "Neutral": 0, "Negative": 0}

    # 3. Process the data
    for video in channel_data:
        print(f"\n--- Analyzing: {video.video_id} ---")

        if not video.comments:
            print("No comments found for this video.")
            continue

        for comment in video.comments:
            # Use comment.text as defined in your Comment model
            scores = analyzer.polarity_scores(comment.text)
            compound = scores['compound']

            if compound >= 0.05:
                sentiment = "Positive"
            elif compound <= -0.05:
                sentiment = "Negative"
            else:
                sentiment = "Neutral"

            stats[sentiment] += 1
            print(f"[{sentiment}] {comment.author}: {comment.text[:50]}...")

    # 4. Final Summary
    print("\n" + "="*30)
    print(f"OVERALL SENTIMENT FOR @{handle}")
    for label, count in stats.items():
        print(f"{label}: {count}")

if __name__ == "__main__":
    analyze_comments("TheOffice")

The AI-Ready Content Summarizer

This script fetches a video's transcript, strips out unnecessary timing info, and combines it with metadata to create a perfectly formatted .txt file that you can drop into any AI tool.

from ytfetcher import YTFetcher
from ytfetcher.services import TXTExporter

def prepare_for_ai(video_ids: list[str]):
    # 1. Initialize using the video URL factory
    fetcher = YTFetcher.from_video_ids(video_ids=video_ids)

    print(f"Downloading transcript for AI processing...")
    channel_data = fetcher.fetch_youtube_data()

    # 2. Extract the text only for a clean 'context' string
    full_transcript = ""
    for video in channel_data:
        # We join the transcript segments into one clean block of text
        full_transcript = " ".join([t.text for t in video.transcripts])
        title = video.metadata.title if video.metadata else "Unknown Title"

    # 3. Use your TXTExporter with specific settings 
    # We turn off 'timing' to make it readable for an AI
    exporter = TXTExporter(
        channel_data=channel_data,
        filename="ai_context_ready",
        timing=False, 
        allowed_metadata_list=['title', 'url', 'description']
    )
    exporter.write()

    print(f"--- PREVIEW FOR AI ---")
    print(f"Title: {title}")
    print(f"Transcript Length: {len(full_transcript)} characters")
    print(f"File saved to: ./exports/ai_context_ready.txt")

if __name__ == "__main__":
    url = ['NKnZYvZA7w4']
    prepare_for_ai(url)