Vigilante

Inspiration

Most of our team can agree that, at some point in our lives, we’ve either experienced firsthand or heard from someone close to us about a situation where our safety or well-being was at risk, yet help didn’t arrive fast enough. The harsh reality is that, even in 2025, Mexico remains one of the most insecure places in the world. Every day, countless people find themselves in dangerous situations, desperately needing immediate assistance. But civil forces are human too, they can’t be everywhere at once. On top of that, emergency services like 911 receive an overwhelming number of calls, making it difficult to distinguish between real threats and false alarms. One day, we gathered as a team and shared our own experiences, reflecting on these challenges. We looked at how countries with a high quality of life handle similar situations and asked ourselves: while we might not be able to fix the root cause, could we create a better way to manage it? And that’s how Vigilante was born. The name of our application came from the etymologic origin of the word Vigilante which is vigilare, meaning “to keep awake”, this term was first used to describe community groups that protected their neighborhood by keeping groups of patrols during the day. And that’s what this whole project pretends to provide, a sense of companionship, community and security.

What it does

• Incident alert generation: Emergencies can occur both in public spaces and private residences. Our system leverages pattern recognition and artificial intelligence to quickly detect these events, notify the appropriate authorities, and alert individuals near the affected area.
• Detection of Standard Distress Signals: This includes cases where a person experiences a medical emergency, faints, or faces any situation requiring immediate assistance.
• Data analysis on high-risk Areas: We analyze crime-prone zones, peak incident hours, and patterns of targeted offenses. Additionally, we assess the most common crimes in specific areas. To streamline reporting, we utilize Gemini to generate monthly reports based on the collected data.
• Optimized emergency response coordination: Our system aims to track the real-time location of patrol cars, ambulances, and other emergency response units to ensure they are aware of incidents nearby. This helps prevent delays caused by a lack of awareness or excessive distance, improving response efficiency and saving lives. ## How we built it Our project is structured into three key components: front-end, back-end, and AI models, each playing a crucial role in delivering a seamless and intelligent security alert system. For the front-end, we built the interface using Next.js, a powerful React framework known for its performance and server-side rendering capabilities. To enhance UI and streamline development, we leveraged ShadCN, a component library that provided pre-styled, customizable elements. Additionally, we used Figma for minor design implementations, ensuring a clean and intuitive user experience. On the back-end, we utilized Python with FastAPI, a modern, high-performance web framework designed for building APIs efficiently. For data management, we implemented MongoDB, a NoSQL database, and MongoDB Atlas, its cloud-based solution, to handle and store large volumes of structured and unstructured data. To facilitate seamless communication between the database and the API, we integrated pymongo, a MongoDB driver for Python, and Unicorn, a tool for managing asynchronous web applications. Finally, the AI and machine learning component was powered by Gemini AI and Vertex AI, two advanced platforms designed for training and deploying ML models. These models were specifically trained to analyze footage and distinguish between violent and non-violent situations using publicly available datasets. Their implementation allowed for real-time threat detection, enhancing the overall effectiveness of the security system. By combining these technologies, we created a robust and intelligent platform capable of delivering real-time alerts, efficient data processing, and AI-driven threat analysis to improve public safety. ## Challenges we ran into At the outset, this was an incredibly ambitious project with a broad scope and high expectations. We had envisioned numerous features and functionalities, but as development progressed, we realized the need to reorganize our priorities and refine our approach. Throughout the process, we had to reevaluate our goals, streamline our efforts, and make strategic decisions to ensure the project's success. One of the biggest challenges we faced was working with technologies that were entirely new to us. Navigating unfamiliar tools, frameworks, and integrations required extensive research, trial and error, and continuous adaptation. Despite these obstacles, we remained committed to overcoming each hurdle, ultimately expanding our technical expertise and strengthening our problem-solving abilities. This journey was filled with valuable learning experiences, and while it demanded significant adjustments along the way, it allowed us to build a more refined and effective solution than we initially imagined. ## Accomplishments that we're proud of Despite all the challenges we faced, we successfully delivered a well-executed and meaningful project one that every member of our team is incredibly proud of. Achieving this was no small feat; it required an immense amount of collaboration, dedication, and perseverance. In the final stretch, sleep became a luxury. Many of us worked tirelessly, pushing through exhaustion to ensure everything came together as planned. Some team members even had other work commitments that couldn’t be postponed, forcing them to juggle multiple responsibilities like true professional multitaskers. Ultimately, it was our strong communication, teamwork, and unwavering commitment that gave this project—and this entire experience—its true meaning. The collective effort and shared determination to see it through made all the difference, transforming challenges into a rewarding and unforgettable achievement. ## What we learned Throughout this project, we worked with technologies that were entirely new to us, pushing ourselves beyond our comfort zones. For most of our team, this was our first experience developing back-end services with Python, as well as our introduction to Machine Learning models and services. Navigating these unfamiliar tools required quick adaptation, problem-solving, and teamwork. Beyond the technical challenges, this experience also taught us a lot about ourselves. We discovered how each of us handles high-pressure situations, collaborate under stress, and finds creative solutions to complex problems. This challenge was not just about building a functional product, it was a test of time management, resource allocation, and our ability to learn and adapt quickly. In the end, this journey proved that with determination, teamwork, and a willingness to learn, we can tackle even the most ambitious challenges. ## What's next for Vigilante We are committed to ensuring the continuity and long-term development of this project, as we firmly believe it has the potential to bring significant value to our country. The increasing trust and familiarity with technology, accelerated by the pandemic, has paved the way for seamless integration between innovation and public services. We envision a future where technology is not just a tool but a harmonious part of every community, enhancing safety, efficiency, and accessibility. Our goal is to fully realize the original vision of the project—expanding beyond security alerts to encompass a comprehensive public service management system. This includes integrating our platform into essential services, from transportation optimization to advanced healthcare monitoring and emergency response coordination. By continuing to develop and refine our solution, we aim to create a smarter, more connected, and safer society, where technology empowers individuals and strengthens public infrastructure. This is just the beginning, and we are excited to push the boundaries of what’s possible.

Built With

• chadcn
• figma
• gemini
• github
• machine-learning
• nextjs
• python
• typescript
• vertexai