TeamSync

Inspiration

This project was born from our shared experiences as freelancers. We often found ourselves in situations where we had to go back and forth with clients just to re-explain or schedule our weekly check-in meetings. While these meetings are helpful, finding a time slot that works for everyone is often a hassle in itself. We wanted to build something that made this process clearer—providing visuals, guides, and descriptions to make the client update loop transparent, fast, and seamless for everyone involved.

What it does

TeamSync is an AI-powered client portal designed to help development teams keep stakeholders informed effortlessly. It automatically generates video and PDF updates when GitHub pull requests are merged, using AI to summarize code changes, add voice narration, and capture relevant visuals. With an integrated AI assistant, role-based access, and project dashboards, TeamSync streamlines communication, collaboration, and progress tracking—all in one platform.

How we built it

We started by sitting down to carefully plan the features and the overall system architecture.
We built the base front-end first to establish a solid visual direction.
We split into two sub-teams (two focusing on Front-End, two on Back-End) to delegate tasks effectively.
We tackled our task list one by one, constantly jumping in to help each other debug whenever someone got stuck.

Challenges we ran into

GitHub Integration: getting the webhooks to trigger correctly was trickier than expected.
Video Generation: automating the video creation process presented several technical hurdles.
Document Generation: We initially planned to use Google Docs, but realized the API costs/paywalls were a barrier. We had to pivot quickly and find an alternative solution using PDF generation instead.

Accomplishments that we're proud of

We are honestly surprised by how much we were able to build. Initially, we just planned to create a simple video automation tool. However, it evolved into a comprehensive system where an AI agent handles automation and personalization using Context Engineering and RAG systems. The whole team worked through the 24 hours non-stop, and we are really proud to have reached a high-quality, full-stack portal that went far beyond our original scope.

What's next for TeamSync

Mobile App: shifting from web-only to an app to enable direct push notifications.
Smarter Meetings: integrating a meeting minute-taker AI that is "context-aware"—meaning it understands the specific project scope and documents to provide better summaries.

What we learnt

We learned that bridging real-world data with artificial intelligence is genuinely difficult because of the risk of hallucinations. To solve this, we had to implement grounding systems and hub integrations to back up our evidence, ensuring the LLM remained reliable. This was also our first time implementing a full AI Agent architecture. It forced us to think not just as coders, but as system designers. It was a huge challenge to map out that timing and structure, but we found it incredibly fun. Beyond the tech, we really learned the value of tight collaboration under pressure.