AccessMate

Inspiration

AccessMate was inspired by the difficulty many users face when navigating complex and information-dense websites. Many individuals with cognitive, reading, or sensory accessibility needs struggle to process cluttered webpages and long unstructured content. The goal of this project was to build a tool that helps make online information easier to understand and more independently accessible.

What it does

AccessMate is an adaptive web accessibility companion that simplifies webpage content by generating concise summaries from user-provided website links. Users can paste a webpage URL, and the system automatically extracts the main article content and processes it using a natural language summarization model to produce readable and simplified output. It also has an accessibility mode for easier viewing in dark colours.

How I built it

AccessMate was built using: • Flask as the backend web framework • Transformers library for AI-based text summarization • BeautifulSoup for webpage content extraction • HTML, CSS, and JavaScript for frontend interface

The summarization model extracts webpage text, removes noisy metadata, and applies a transformer-based sequence-to-sequence model to generate simplified summaries.

Challenges I ran into

Challenges included handling noisy webpage structures during scraping, managing model inference latency in a cloud development environment, and preventing fragmented summarization outputs. I addressed these by optimizing extraction logic and tuning generation parameters.

Accomplishments that I’m proud of

I am proud of developing a functional prototype that demonstrates the potential of adaptive accessibility technology. AccessMate integrates web scraping, machine learning summarization, and user interface interaction into a single tool aimed at improving digital accessibility.

What I learned

This project helped me understand practical AI integration in web applications, frontend-backend communication, and accessibility-focused software design. I also learned about limitations of local inference models and optimization trade-offs.

What’s next for AccessMate

Future improvements may include voice-based accessibility assistance, personalized cognitive adaptation modes, browser extension deployment, and scalable cloud inference for improved summarization quality.

Built With

• beautiful-soup
• css
• flask
• github
• html
• huggingface
• javascript
• python
• pytorch
• transformers