TranscriptIQ is a project that enables users to transcribe YouTube videos and perform various NLP (Natural Language Processing) tasks, chat with youtube video and many more on the transcribed text.
Streamlit.LLM.mp4
./
├── LICENSE
├── README.md
├── app.py
├── images
│ ├── AI.jpg
│ ├── NER.png
│ ├── Robot_whisper1.jpg
│ ├── Robot_whisper2.jpg
│ ├── Robot_whisper3.jpg
│ ├── expander.png
│ ├── mp3_2_text.png
│ ├── resum.png
│ └── youtube2text.png
├── myfunctions
│ ├── __pycache__
│ ├── my_functions.py
│ └── my_summarization_functions.py
├── packages.txt
├── pages
│ ├── Chat.py
│ ├── Transcribe Youtube.py
│ └── static
└── requirements.txt
6 directories, 18 files
The main functionality of TranscriptIQ is provided by the transcribe_youtube.py script. It uses the Streamlit library to create a user interface for transcribing YouTube videos and performing NLP tasks. Here is a brief overview of the functionality provided by the script:
- Retrieve YouTube video information (title, author, views, duration, etc.)
- Download YouTube videos as MP4 format
- Convert MP4 videos to MP3 audio files
- Transcribe MP3 audio files to text using speech-to-text technology
- Perform text summarization using the Cohere API
- Perform Named Entity Recognition (NER) using the Spacy library
- Generate graphs based on NER results
- Perform sentiment analysis using the VADER library
- Display word clouds based on NER results
To use TranscriptIQ, you can run the app.py script. This will start the Streamlit app and you can interact with it to transcribe YouTube videos and perform NLP tasks on the transcribed text.
Note: Before running the script, make sure you have installed all the required packages listed in requirements.txt.
streamlit run app.pyTranscriptIQ was developed as part of the Streamlit LLM Hackathon. The project was created by Devanshu, Somesh.
- demo running
- understand pricing points on customgpt.ai
- plan for a the link of videos. ( 20 videos downloaded as mp4 )
- build a demo according to the plan.