EchoTrace - The AI That Finds the Missing Context

Inspiration

Have you ever opened an old document, email, or screenshot and thought, “Why did I save this?” or “What was I working on?” We noticed that our digital lives contain a lot of small pieces of information, but they are often disconnected. An email may be related to a document, a calendar event, or even a location we searched days ago. We wanted to build something that could connect these pieces and help people recover the context they had forgotten.

What We Built

We built EchoTrace, an AI-powered tool that connects different parts of your digital activity and tries to understand the story behind them. Users can connect sources such as emails, documents, calendar events, and locations. EchoTrace looks for relationships between these pieces and uses machine learning and Gemini to understand what they might be connected to. For example, if you received an email about an interview, saved a related document, added an interview to your calendar, and searched for the company location, EchoTrace can connect these clues and tell you:

“You were preparing for an interview with this company.” It also shows the information that led it to that conclusion, so the user can understand why the AI made that connection.

How We Built It

We use Google APIs to access relevant user information such as emails, calendar events, documents, and locations. We use machine learning to identify relationships between different pieces of information. Gemini is then used to understand the collected context, connect the different clues, and explain the most likely story in simple language. The main flow is: User Data → Extract Information → Find Connections → Build Context → Gemini → Reconstructed Story

What We Learned

While building EchoTrace, we learned that making an AI system useful is not just about generating a good answer. It is also about giving the AI the right context and showing users why an answer was generated. We also learned how different Google APIs can work together to create something much more useful than a standalone AI chatbot.

Challenges

One of our biggest challenges was deciding which connections between different pieces of information were actually meaningful. Not every email, document, or calendar event is related. We also had to think carefully about privacy because EchoTrace works with personal information. We designed the concept around user control, transparency, and showing the evidence behind AI-generated conclusions.

The Goal

Our goal with EchoTrace is simple: Help people find the context they lost, instead of making them search for it all over again.

Built With

• fastapi
• gemini-api
• google
• google-maps-platform
• mongodb
• next.js
• python
• react
• scikit-learn
• typescript