EduAccessAI

Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for EduAccessAI

EduAccess AI

Inspiration

Inclusive education remains a major challenge worldwide, particularly for students with disabilities who face physical barriers in accessing learning environments. Through secondary research from UNESCO, WHO, UNICEF, and the World Bank, I discovered that many schools still lack accessible infrastructure such as ramps, clear pathways, accessible entrances, and disability-friendly facilities.

Traditional accessibility audits often require trained specialists, site visits, and significant costs, making them difficult for many schools to conduct regularly. This inspired me to explore how artificial intelligence could make accessibility assessments faster, more affordable, and more accessible.

What it does

EduAccess AI is an AI-powered accessibility auditing platform that analyzes images of school environments and evaluates how accessible they are for people with disabilities.

Users simply upload a photo of a school environment, and the system:

Generates an Accessibility Score (0–100)
Identifies accessibility barriers
Highlights positive accessibility features
Provides actionable recommendations
Explains the reasoning behind the score
Generates downloadable PDF reports
Converts reports into audio using Text-to-Speech for visually impaired users

The goal is to help schools quickly identify areas for improvement and support inclusive education planning.

How I built it

EduAccess AI was built using:

Python
Streamlit
Google Gemini AI
Pillow
ReportLab
gTTS (Google Text-to-Speech)
python-dotenv

Google Gemini provides the image understanding and reasoning capabilities that power the accessibility assessments. Streamlit was used to create an intuitive dashboard where users can upload images, view reports, listen to audio summaries, and download PDF reports.

Challenges I ran into

One of the biggest challenges was configuring and connecting the Gemini API correctly. I encountered model compatibility issues, package version conflicts, and environment configuration challenges before successfully integrating image analysis.

Another challenge was designing a structured reporting format that produced consistent outputs suitable for scoring, recommendations, PDF generation, and text-to-speech conversion.

What I learned

Through this project, I gained practical experience working with multimodal AI systems, prompt engineering, accessibility-focused design, Streamlit application development, API integration, PDF generation, and assistive technologies such as text-to-speech.

I also learned the importance of designing technology that is itself accessible and inclusive.

What's next for EduAccess AI

Future versions will include: