Identified Problem: When working on group projects, students often struggle with communication. Members lose track of their assigned tasks, putting off work for others to do. Group chats, which are intended to simplify collaboration, can instead become cluttered with irrelevant messages, making it difficult to stay updated on project progress. This frequently leads to miscommunication, uneven workload distribution, and last-minute cramming.
Solution: To address this, we developed a web application that streamlines communication, planning, and task management within group projects. The platform features a live chat for real-time collaboration. However, beyond that, after discussions take place, the chat can be automatically summarized to highlight key messages, decisions, and important information. Alongside this, the software generates a visual workflow diagram that outlines the group’s agreed-upon plan. Users can then upload a rubric or list of deliverables. The software analyzes the rubric and extracts the key requirements, presenting them in a clear and organized format so users understand exactly what is needed for full completion. As the group begins working, members can continuously upload updated versions of their project while noting what changes or contributions they made. The software analyzes each new version and compares it against the rubric, visually indicating which requirements have been met and which have not. Based on this analysis, the platform automatically generates a dynamic to-do list that outlines the remaining tasks needed to complete the project. Each task includes estimated completion times, helping users better manage their workload and stay on schedule. This ensures that all members remain aligned, accountable, and aware of overall progress.
Application Use: Users begin by communicating through the live chat to establish a plan. Regardless of who is available, the chat is summarized and a workflow diagram is generated to clearly outline the project structure. Even without the chat, any user on the platform knows exactly what to do with no communication because of the clear “small task” assignment feature, along with how much of the total work is left. Next, users upload the project rubric, which the software breaks down into key requirements. As work progresses, members upload updated versions of the project and note their contributions. With each submission, the system evaluates the project against the rubric, visually tracking completed and incomplete requirements. The software then updates a shared to-do list, including specific tasks and estimated completion times, guiding the team toward full completion. This continuous feedback loop helps ensure the project stays organized, efficient, and aligned with all grading criteria.
Technical Overview: The application was made using nodejs and socketio in combination with a backend of Google’s Gemini 3.1 Flash Lite API. We have 2 main AI powered features. The first is used to generate the summary of the chat. People coming back to hundreds of messages may be overwhelmed, and this feature summarizes the information relevant to the project using prompt engineering with Gemini, and displays it in bullet points and a mermaid chart using mermaid.js The second allows us to upload a rubric and compare versions of the project to it, and create quick deliverables to ensure anyone can contribute at any time, regardless of how much time they have. NodeJS allows for instant updates, and Gemini allows for flexible, project specific feedback. By only using information contained within the documents, along with prompt engineering to ensure brief, specific feedback, the risk of hallucination is minimized. In addition, these updates are linked to the rubric, showing which points are addressed, and which are missing, along with a completion bar to give members progress at a glance.
Log in or sign up for Devpost to join the conversation.