Inspiration
Many students don’t have a reliable way to track their actual grade. Some professors don’t provide a running total, and when they do, it’s often outdated or inaccurate. We wanted to build something that gives students clarity and control over their academic performance, without relying on inconsistent systems.
What it does
Grade Harvester allows students to upload their course syllabus and automatically extracts the grading breakdown using AI. It also includes a Chrome extension that securely pulls in assignment scores from learning platforms. Together, this enables accurate, real-time grade calculations, even in courses where no official grade is provided. It also handles complex rules like dropped assignments, weighted categories, and extra credit, so students always know where they stand.
How we built it
We built the frontend using React and Material UI for a clean, intuitive interface. The backend was developed in Go to handle file uploads, processing, and API communication. When a syllabus is uploaded, we extract the text and send it to the Gemini API, which parses grading rules into structured data.
We also built a Chrome extension that reads assignment and grade data directly from supported learning platforms and sends it to our backend, allowing us to combine real performance data with extracted grading policies for accurate calculations.
Challenges we ran into
One of the biggest challenges was handling the variability of syllabi. Every professor formats grading policies differently, and some include complex rules like dropping the lowest scores or conditional extra credit. Ensuring consistent and structured outputs from the AI was also difficult.
On the extension side, adapting to different page structures across learning platforms and reliably extracting grade data was another major challenge.
Accomplishments that we're proud of
We’re proud of building a system that turns unstructured syllabi into actionable grading systems and combines that with real assignment data. The integration of AI parsing with live grade inputs creates a level of accuracy and automation that most students don’t currently have access to.
We’re also proud of successfully building both a full-stack app and a browser extension within a short timeframe.
What we learned
We learned how to integrate AI into a real-world pipeline and the importance of enforcing structured outputs. We gained experience building a Go backend, handling file parsing, and developing a Chrome extension that interacts with live web content.
We also learned how to design around messy, real-world data instead of ideal inputs.
What's next for Grade Harvester
Next, we want to expand support for more learning platforms and make the extension more robust across different course layouts. We also plan to add predictive features, like telling students what they need on upcoming assignments to reach a target grade.
Long term, we want to turn Grade Harvester into a complete academic dashboard that gives students full visibility and control over their performance.
Log in or sign up for Devpost to join the conversation.