AccessiCheck

Please check the slide deck in the link for more information and for our full pitch video!

✨ Inspiration

Our teammate Tasnia works at the Ministry of Seniors and Accessibility, where she helps ensure that government websites comply with Accessibility for Ontarians with Disabilities Act (AODA) standards. Her real-world co-op experience helped us understand what accessibility barriers Ontarians face, inspiring us to build a tool that makes PDF accessibility simpler and scalable.

❓ What it does

PDFs are everywhere—from filling out mortgage applications to gathering requirements from legal documents. While these files are technically available, they are not always accessible. This creates barriers for seniors, neurodivergent users, newcomers, ESL learners, and anyone relying on assistive technologies to navigate essential information. In Ontario, the AODA requires that important government files are accessible under an Accessibility Evaluation Framework. For example, font sizes must be at least 12pt and sentences must be short (average of less than 20-25 words). However, ensuring compliance is time-consuming, inconsistent, and often overlooked, where many important PDFs still fail to meet accessibility standards, leaving a large group of Ontarians disadvantaged.

AccessiCheck aims to address this gap. Users can upload any PDF, and AccessiCheck analyzes the document against AODA and Web Content Accessibility Guidelines (WCAG) 2.1 AA criteria. It then generates: • A compliance score • An audit report breaking down accessibility issues and points for improvement (categories include contrast, hierarchy, typography, layout, alt text, etc.)

AccessiCheck also goes a step further: it automatically edits the PDF to fix issues and ensure accessibility. This version is available for download (includes simplified text, semantic structure, alt text for images, high-contrast styling, and audio for TTS), allowing organizations, government bodies, and everyday users—especially those who need it the most—to produce inclusive and accessible PDFs instantly.

💻 How we built it

Frontend: Simple upload interface for PDFs or images displays score, suggestions, audio, and an accessible version of the uploaded pdf once ready. Backend: Node/Express server handles file uploads, extracts text/images from PDFs, calls Gemini AI, and returns results.

Processing Flow: Extract document content (text + images) Send content to Gemini LLM with AODA/WCAG 2.1 AA evaluation checklist (from public domain) embedded in prompt Receive JSON response: score (% adherence to AODA guidelines), breakdown, and suggestions Generate corrected document (text/HTML) of the inputted PDF

Accessibility Features: Simplified text, semantic structure, alt text for images, high-contrast styling, downloadable PDF + audio for TTS

🏃‍♀️ Challenges we ran into

We were new to using the Google Gemini API, so we had to learn how to implement it and produce a viable prototype effectively in the short timeframe we were given.

❗ Accomplishments that we're proud of

Despite the time limit of a single-day hackathon, we are proud that we were able to produce a working prototype that has the possibility of making a social impact.

➡️ What's next for AccessiCheck

We want to expand AccessiCheck to further enhance accessibility and usability. We want to extend to full website scanning, more standards, multi-language support, and live updates.