A game-changing web application that transforms the way you experience online video content.
ClipInsight.1.mp4
ClipInsight is the result of a creative whirlwind during LLM Hackathon by Streamlit and Generative AI hackathon by Peerlist, where we set out to tackle the frustration of navigating through lengthy video content on platforms like YouTube. Our mission was to make video content more accessible, interactive, and time-efficient for users from all walks of life.
Why? Because we believe that knowledge should be easily accessible and engaging for everyone.
-
📋 Text-Based Summaries with Timestamps: Our app analyzes videos and generates concise summaries of key moments, complete with timestamps. No more endless scrolling!
-
💬 Real-Time Chat: Engage in dynamic discussions while interacting with our AI assistant. Ask questions and share insights.
- Python
- Streamlit
- Langchain
- AssemblyAI
- Replicate
- Model Used: LLama2-70B
To get started with ClipInsight, follow these simple steps:
-
Fork this repository
-
Clone the repository:
git clone https://github.com/yourusername/your-repo.git -
Install dependencies:
pip install -r requirements.txt -
Run the app:
streamlit run app.py
We welcome contributions from the community. Whether it's code, design, or ideas, your input is valuable.
I have added 2 unit tests described below :-
- download_video_test.py - this unit test tests the download_video() method which actually extracts audio from a youtube video url.
- split_transcript_test.py - this unit test tests the split_transcript() method which is a utility method in itself and helps in breaking a huge file of text into several chunks.
- To Run the test file:
python -m unittest <file_name>.py
Checkout this Commit for more details.