HackMIT 2025 Project

https://plume.hackmit.org/project/hictt-pbcbv-uzktl-zgrym

Inspiration

We were inspired by witnessing firsthand how teams struggle with AI collaboration in existing platforms. Through our market research, we discovered critical pain points: teams duplicating AI work, experiencing misalignment due to inconsistent AI responses, and losing valuable context because AI tools operate in silos. Stories like "AI edits files it isn't supposed to and messes up documentation" and teams getting "parallel instructions not the same ones" highlighted a massive gap in how current platforms handle AI integration. We saw an opportunity to transform team collaboration by creating an intelligent workspace that learns collectively rather than forcing teams to start from scratch with each AI interaction.

What it does

Jova is a collaborative AI workspace that seamlessly integrates messaging, file storage, and AI agents into one intelligent platform. Our solution addresses three core problems:

1. Collaborative Server with Integrated AI: Combines messaging, video calls, and AI agents in a unified workspace, eliminating the need to switch between multiple platforms.
2. Shared AI Context & Knowledge Building: AI agents learn from collective team knowledge, including previous AI conversations and shared files, ensuring consistency and reducing redundant work.
3. Smart File Storage & Retrieval: Enables teams to find files by content rather than filename, making important documents and insights easily discoverable even when exact names are forgotten.

The platform transforms how teams work together by creating a persistent, intelligent memory that grows with team collaboration.

How we built it

Backend Infrastructure: Developed a collaborative server architecture that handles real-time messaging, video calls, and AI agent coordination AI Integration Layer: Created shared AI context management that allows agents to learn from collective team interactions and file contents Smart Storage System: Implemented content-based file retrieval using advanced search algorithms that index file contents rather than just metadata User Interface: Designed an intuitive workspace that seamlessly blends traditional collaboration tools with AI capabilities Knowledge Management: Built a system that captures and retains context from AI conversations and file interactions to inform future team decisions

Built With

• ai-agents
• python