SafetySphere

• 

  SafetySpher, Digital AI Bodyguard

Inspiration

SafetySphere was inspired by the urgent need for accessible, real-time safety information and responsive reporting tools within communities. Events where delays or lack of detail in alerts resulted in serious consequences showed us that existing solutions are fragmented. Our motivation was to build a platform where users can not only receive timely safety alerts but also proactively share incident reports, missing persons, or pets, all supported by smart, AI-driven guidance.

What it does

SafetySphere is an AI-powered safety assistant app that delivers real-time alerts, enables users to file and track incident reports (such as theft, crime, missing persons, or pets), and provides practical guidance via a multimodal chatbot. By integrating live data from sources like BrightData, SafetySphere aggregates localized safety updates. The app’s NLX-powered chatbot allows users to interact naturally with text, photos, or videos—helping them file reports, identify missing individuals or pets, and access immediate advice. SafetySphere also offers short, context-aware safety videos generated by Heygen, and securely stores all data for easy retrieval and analytics.

How we built it

Data Aggregation: We connected BrightData APIs to fetch real-time safety information (e.g., crime reports, alerts, missing persons and pets) from diverse open data sources. Multimodal Chatbot: Integrated the NLX chat agent for natural, intuitive user interactions, supporting not only text but also processing of user-uploaded photos and short videos. Video Integration: Leveraged Heygen to generate and embed targeted, safety-oriented instructional videos based on the alert or reporting context. Backend/Data Storage: Used MongoDB for secure, scalable storage of user submissions, chat logs, and all associated media. Frontend/UI: Designed an intuitive interface for alert consumption, report submission, video viewing, and archival search—all optimized for speed and usability under stress.

Challenges we ran into

Data Standardization: Aggregating various safety data feeds with differing formats and levels of reliability required substantial normalization and filtering to ensure accuracy.

Accomplishments that we're proud of

Successfully combined multimodal AI (chat, image) into a unified, user-friendly application. Achieved real-time aggregation and delivery of safety alerts personalized to a user’s current location. Built seamless reporting flows that allow community members to participate actively in public safety. Established a secure, scalable infrastructure to store.

What we learned

Multimodal interfaces (supporting text, images) vastly increase accessibility and effectiveness in safety-focused applications. Real-time community safety is as much about fast, trustworthy information flow as it is about technology—users need tools they can trust and use under pressure. Effective data aggregation demands robust normalization pipelines and continuous quality checks. Privacy concerns are paramount; user trust relies on transparent, secure data handling and communication.

What's next for SafetySphere

Expand data coverage internationally and add more real-time public source integrations. Enrich NLX chatbot capabilities to handle more languages and emergency scenarios. Enhance image and video recognition with advanced AI models for even faster, more accurate identification. Pursue partnerships with local, state, and national agencies to ensure broader alert coverage and reporting reach. Launch SafetySphere on additional platforms ( web dashboard) to maximize accessibility.

Built With

• brightdata
• heygen
• mongodb
• nlx
• trae