Summarizer X

Inspiration

Accessing and processing digital information can be challenging for individuals with visual impairments, while sighted users require tools for efficient text processing. A unified platform that provides intuitive solutions for summarization, rephrasing, and translation can bridge this gap, empowering both groups to interact with information effortlessly and inclusively.

What it does

We are developing an inclusive, multi-functional website designed to cater to both blind and normal users by providing access to advanced text summarization, rephrasing, and translation features. The core goal of this solution is to enhance accessibility, ease of use, and learning for all users, with a special focus on supporting blind individuals.

How we built it

We have used the below mentioned tech stack to build this, Language Model for Blind Users: LLama LLM (Large Language Model) Text-to-Speech (TTS) Web Technologies for User Interface: Frontend: HTML, CSS, and JavaScript Backend: Flask or Django (depending on preference) PDF Generation: pdfkit Word Document Generation: The docx library Data Sources and APIs : Wikipedia API Groq API Google Translate API

Challenges we ran into

Voice Recognition Accuracy: Challenge: Ensuring that the system can accurately understand voice commands from blind users, even in noisy environments. Solution: We implemented a dynamic energy threshold and adjusted the sensitivity of the microphone to reduce the impact of background noise. Multiple retries were allowed for better accuracy.

1. Text-to-Speech (TTS) Processing: Challenge: The quality of TTS could sometimes be robotic or unclear, especially for long sentences or complex words. Solution: We integrated pyttsx3 for TTS, providing support for different voices and languages. We also split longer text into smaller chunks to ensure better readability when speaking aloud.
2. Handling Large Content: Challenge: Processing large documents or articles, particularly when dealing with summaries, could result in memory overload or long processing times. Solution: We truncated large content before processing it through the LLM (using the truncate_content function) to ensure efficiency and minimize delays.

Accomplishments that we're proud of

We had faced few challenges while building this and we have managed bring the suitable solution for those problem.

What we learned

In this project we have learned about groq API and LLM model

What's next for Summarizer X

1. Expanding translation capabilities to include more languages, dialects, and real-time speech translation will make the tool accessible to a global audience.
2. Creating dedicated apps for Android and iOS with accessibility features like screen readers and gesture-based navigation will ensure usability for both visually impaired and general users.
3. Integrating machine learning to provide tailored summaries and rephrased content based on user preferences, reading levels, and learning styles will enhance the user experience significantly.

Built With

• css
• groq
• html
• llm
• python
• streamlit
• tts