In the digital age, video content has become a primary source of information, yet extracting specific details from lengthy YouTube videos remains a challenge. QueryVid aims to solve this by providing a seamless interface for querying YouTube video content. By leveraging advanced APIs for video transcription and natural language processing, QueryVid allows users to obtain precise answers to their questions based on the content of the video, thereby enhancing information accessibility and efficiency for developers and users alike.
To run QueryVid on your local machine, follow these steps:
-
Ensure you have the latest version of Python and pip installed.
-
Clone the repository:
git clone https://github.com/aniiishhh/QueryVid.git cd QueryVid -
Create and activate a Python virtual environment:
python -m venv venv source venv/bin/activate # On Windows use `venv\Scripts\activate`
-
Install the required packages:
pip install -r requirements.txt
-
Get your API keys:
-
Host a demo PostgreSQL database locally or on cloud services (suggested: Render).
-
Enter the database details and the API keys in the
.envfile in the root directory. Your.envfile should look something like this:ASSEMBLYAI_API_KEY = your_assemblyai_api_key OPENAI_API_KEY = your_openai_api_key DB_ENGINE = "django.db.backends.postgresql_psycopg2" DB_NAME = your_db_name DB_USER = your_db_user DB_PASSWORD = your_db_password DB_HOST = your db_host DB_PORT = your db_port -
Finally, run the application:
python manage.py runserver
Now you can open your web browser and navigate to http://127.0.0.1:8000/ to try out the app.
Frontend: HTML, Tailwind CSS, JavaScript
Backend: Python, Django
Database: PostgreSQL
APIs: AssemblyAI Speech-to-Text API, OpenAI GPT-3.5-turbo API
QueryVid is designed to provide a user-friendly interface for querying YouTube video content. Users can sign up and log in to access the main features of QueryVid, ensuring that their queries and history are securely stored and accessible only to them. Once logged in, users enter a valid YouTube video link and a question into the provided input fields. The backend processes this request by extracting video details and downloading the audio file using the yt_dlp library.
The audio file is then transcribed into text using the AssemblyAI API. This transcription is sent to the OpenAI GPT-3.5-turbo model along with the user's question. The model generates an answer based on the video content. If the task is successful, the query details and results are stored in a PostgreSQL database for future reference, accessible in the query history section.
QueryVid's frontend is crafted with HTML and Tailwind CSS, providing a responsive and visually appealing user experience. JavaScript is used to enhance interactivity and ensure smooth transitions and animations. The backend is powered by Django, which handles the business logic and integrates seamlessly with the PostgreSQL database to manage user data and query history efficiently.
Robust error handling is incorporated to manage invalid YouTube links, transcription errors, and issues with the OpenAI API. Users receive clear messages to guide them in correcting their input or understanding the error. This ensures a smooth user experience and helps maintain the application's reliability.
QueryVid is an innovative solution for extracting valuable information from YouTube videos. By combining powerful APIs with a robust backend and a clean, interactive frontend, it delivers a unique and efficient user experience.
The use of yt_dlp for video extraction, AssemblyAI for transcription, and OpenAI's GPT-3.5-turbo for generating answers demonstrates a sophisticated workflow that leverages cutting-edge tools and services.
Feel free to reach out or connect with me on LinkedIn and GitHub for more information or to discuss other projects.