Levi: From Capstone Idea to AppCon Finalist
2025 | Milestone | Best in Thesis
.jpg)
.jpg)
.jpg)
.jpg)
Innovation often begins with a simple question. For our team, that question was: "Why is legal information so difficult for ordinary Filipinos to understand?" Laws exist to protect citizens, yet many people struggle to access or interpret them due to complex legal language, lack of awareness, and limited access to legal professionals. From that challenge, Levi was born.
The Beginning of Levi
Levi stands for Legal Empowerment Through Artificial Intelligence, inspired by the biblical book of laws, Leviticus. Our goal was to build a system that bridges the gap between complex legal systems and everyday understanding in the Philippines.
We wanted to create a platform that could simplify legal information, provide reliable references, and help people understand their rights and responsibilities. Rather than replacing lawyers, Levi was designed to serve as an educational assistant—a tool that empowers citizens with accessible legal knowledge.
This idea eventually became our Capstone Project, a culmination of months of research, development, and collaboration.
Our Journey to AppCon 2024–2025
With Levi developed as our thesis project, we decided to take it a step further. Our team joined the 2024–2025 AppCon Competition, a national-level event where student innovators showcase technological solutions to real-world problems.
Competing against many strong teams from different schools, we presented Levi as a platform that could transform how Filipinos interact with legal information. Through our demonstrations and presentations, we showed how AI could simplify legal language, answer legal inquiries, and guide users toward credible legal resources.
After several stages of evaluation and judging, our project advanced to the finals. Reaching the final stage of AppCon was a proud moment for our team and validated the potential impact of Levi.
Even more meaningful was the recognition our project received within our academic program. Our team was awarded Best in Thesis, a testament to the dedication and effort we poured into building Levi.
What Levi Does
For the Public
Levi helps citizens understand laws through simplified explanations, multilingual support, and scenario-based examples. This allows users—even those without legal backgrounds—to better understand their rights.
For Legal Practitioners
The platform can assist lawyers by improving legal research efficiency, organizing case references, and supporting client consultation processes.
For Government and Policy Analysis
Levi can analyze laws and identify possible inconsistencies or overlaps within legal frameworks, helping policymakers better understand how laws interact with each other.
By making legal information easier to access and interpret, Levi aims to strengthen civic awareness and legal literacy.
The Technical Landscape of Levi
Behind Levi is a modern technology stack designed to support AI-powered legal assistance.
Next.js
Frontend framework for fast, responsive UI
Vercel
Hosting platform for scalable web performance
Google Cloud Run
RESTful APIs with dynamic scaling
MongoDB Atlas
Database for metadata and embeddings
Google Cloud Storage
Secure storage for legal documents
Gemini AI
Core AI model for response generation
RAG Architecture
Retrieval-augmented generation for accuracy
LangSmith
AI observability and debugging
The frontend is built using Next.js, which provides a fast and responsive web interface for users. The platform is hosted on Vercel, ensuring scalable and reliable web performance.
The backend operates through RESTful APIs deployed on Google Cloud Run, enabling the system to handle requests dynamically and scale based on demand. For data storage, Levi uses MongoDB Atlas to manage metadata, embeddings, and system logs.
Files and legal documents are stored using Google Cloud Storage, which allows the system to securely manage large legal datasets.
At the core of Levi's intelligence is an AI model powered by Gemini, which generates responses and performs semantic reasoning. To ensure that responses remain grounded in real legal references, Levi uses a Retrieval-Augmented Generation (RAG) architecture. This means the system retrieves relevant legal documents and contextual information before generating answers, improving both accuracy and reliability.
To monitor and improve AI behavior, the system uses LangSmith for observability. This allows developers to trace responses, debug prompts, and maintain transparency in how the AI processes legal queries.
Together, these technologies create a platform capable of delivering accessible, contextual, and reliable legal information.
Looking Forward
Levi began as a thesis project, but our journey at AppCon showed us that the idea has potential far beyond the classroom. The challenge of legal literacy in the Philippines remains significant, and technology can play a role in addressing it.
Through Levi, we hope to contribute to a future where every Filipino can easily understand the laws that protect them.
Our experience building Levi taught us more than technical skills—it showed us how technology can be used to empower communities. From a simple concept to a national competition finalist and a Best in Thesis award-winning project, Levi represents our team's passion for innovation and meaningful impact.
And this is only the beginning of the journey.

.jpg)
.jpg)