ytdebunk 1.0.4

A CLI tool to download audio from a YouTube video, transcribe it, and refine the transcription using AI.

ytdebunk

Overview

Current Features:

ytdebunk is a command-line tool designed to:

• Download audio from YouTube videos.
• Transcribe the audio content.
• Optionally enhance the transcription using the Gemini API.
• Optionally detect logical faults in the transctiption using the Gemini API.

Features in queue:

• Classifying assertive claims from the transcription.
• Fact-checking and validation of the claims from reliable source using online search and agentic AI.
• Re-organizing the factual faults and logical faults.
• Preparing a script for a hypothetical debunker charecter using generative AI (or AI Agents).
• Synthesizing the script to create an audio and a video using generative AI (or AI Agents).

This tool is particularly useful for analyzing transcriptions to identify logical fallacies and incorrect claims made by YouTubers and prepare a debunk video.

Installation

For avoiding conflicts better create a virtual environment and start working on it:

python3.11 -m venv .venv
source .venv/bin/activate

Now, you can install from PyPI using,

pip install ytdebunk

Alternatively, for latest updated please try installing directly from Github using:

pip install git+https://github.com/hissain/youtuber-debunked.git

Usage (The CLI Tool)

The ytdebunk is a command-line interface (CLI) with several options.

Arguments

• video_url (str) – URL of the YouTube video to download audio from.

Options

Option	Description
-e, --enhance (bool)	Enhance the transcription using the Gemini API. (Default: False)
-d, --detect (bool)	Detect logical faults in the transcription using Gemini API. (Default: False)
-v, --verbose (bool)	Increase output verbosity.
-t, --token (str)	API token for the Gemini API (Required if --enhance or --detectis enabled).
-st, --start_time (float)	Start time of the audio clip in seconds
-et, --end_time (float)	End time of the audio clip in seconds
-m, --model (str)	A transcription model name from Huggingface (WhisperFeatureExtractor)

Example Usage

ytdebunk "https://www.youtube.com/watch?v=example" -e -d -v -t YOUR_GEMINI_API_TOKEN

export GEMINI_API_TOKEN="your_api_key"
ytdebunk "https://www.youtube.com/watch?v=example" -e -d -v #when Gemini API key is in environment

See an example notebook Example Notebook file for details usage.

Usage (The Streamlit App)

You can simply run the streamlit app to see the demo.

Install the streamlit using pip

pip install streamlit

Run the app.py using streamlit

streamlit run app.py

Screenshots of the Streamlit App

Environment Variables

If preferred, you can set the Gemini API token as an environment variable instead of passing it as a CLI argument:

export GEMINI_API_TOKEN="your_api_key"

Detailed Process

1. Download Audio

   • Uses the download_audio function from ytdebunk.downloader to download audio from the given YouTube URL.

2. Transcribe Audio

   • Uses the transcribe_audio function from ytdebunk.transcriber to generate a text transcription.

3. Enhance Transcription (Optional)

   • If --enhance is enabled, the script uses enhance_transcription from ytdebunk.refiner to refine the transcription using the Gemini API.
   • The API token must be provided via --token or as an environment variable.

4. Detect Logical Faults (Optional)

   • If --detect is enabled, the script uses detect_logical_faults from ytdebunk.philosopher to detect logical fults, fallacies, bias, irony and so on in the transcription using the Gemini API.
   • The API token must be provided via --token or as an environment variable.

5. Save Transcription

   • The final transcription and logical faults (raw or enhanced) are saved to the ./download folder.

Error Handling

• If --enhance or --detect are enabled but no Gemini API token is provided, the script prints an error message and exits.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Contribution and Contact

You can fork this project and submit pull request in the project. Please contact to the author at [e-mail]