Inspiration

Students often never get good teacher feedback or often get it too late to learn anything. This was the inspiration to create a tool that gives advice while the student is still working based on the assignment requirements.

What it does

• Analyzes student assignments against specific rubrics.

• Identifies which parts of the instructions were followed.

• Highlights missing requirements.

• Provides tips to improve the grade before final submission.

How we built it

• Frontend: HTML and CSS for a clean interface.

• File Reading: PDF.js and Mammoth.js to read documents in the browser.

• AI: Google Gemini 1.5 Flash for the analysis.

• Language: Vanilla JavaScript.

Challenges we ran into

• Extracting text from PDFs and Word files without a backend server.

• Managing code to ensure files were read before the AI started.

• Fixing module export errors with the Gemini SDK.

Accomplishments that we are proud of

• Creating a tool that works entirely in the browser.

• Keeping student data private by processing files locally.

• Building a working prototype in a very short amount of time.

What we learned

• How to use the Gemini API for educational feedback.

• Methods for handling document parsing in the frontend.

• The importance of clear prompt engineering for academic tools.

What is next for ClarifyEDU

• Connecting the tool to platforms like Google Classroom.

• Adding support for more file types.

• Creating a dashboard for teachers to see common student mistakes.

Built With

• gemini-api
• html5
• javascript