-
-
Main Home page UI
Inspiration
Public safety reporting is often reactive, fragmented, and inaccessible to many citizens—especially those with disabilities or non-English speakers. Inspired by the need for inclusive, fast, and intelligent civic engagement, we built CivixAI, an AI-powered chatbot that empowers residents of Indianapolis to report safety concerns, get emergency alerts, and access critical services with ease.
What it does
CivixAI is a conversational safety assistant that:
- Allows residents to report incidents (e.g., fire, crime, medical emergencies, hazards).
- Uses AI to extract key details like location, urgency, and incident type—even when users don’t mention them explicitly.
- Provides live weather alerts, emergency contact numbers, and safety tips.
- Supports accessibility needs with simplified responses and inclusive modes.
- Sends real-time notifications to authorities for high-priority incidents.
- Enables media uploads (images/videos) to provide more context for reports.
- Offers an admin dashboard for reviewing incident logs and tracking community trends.
How we built it
- Backend: Flask-based REST API with SQLite for persistent storage.
- AI/NLP: OpenAI's GPT model for intent detection, incident detail extraction, and multi-language support.
- Frontend: Chat interface using HTML/CSS with dynamic JSON-based responses.
- Weather & Alerts: Integrated OpenWeatherMap API for real-time weather hazard notifications.
- Email Alerts: Auto-sends high-priority incidents to city officials using SMTP.
- File Uploads: Supports images, videos, and PDFs for richer incident context.
- Accessibility & Language: Built-in flow adjustments based on user’s needs/preferences.
Challenges we ran into
- Ensuring reliable AI extraction for varied natural language inputs.
- Managing multi-modal input (text + media) without losing performance.
- Balancing data privacy while collecting potentially sensitive incident information.
- Making the system accessible and usable for users with impairments and different languages.
- Handling error cases gracefully when APIs fail or users don’t provide enough detail.
Accomplishments that we're proud of
- Built a fully functional end-to-end incident reporting system in a conversational format.
- Developed robust fallback logic to handle incomplete user inputs.
- Achieved a smooth integration of AI, real-time data, and civic processes.
- Designed with accessibility and inclusivity in mind from day one.
- Created a modular, scalable backend ready for city-wide deployment.
What we learned
- The importance of context-aware AI in public service use cases.
- How to design for civic tech with empathy—prioritizing accessibility, clarity, and trust.
- Strategies for managing secure, real-time data flow between users and authorities.
- That citizens are more likely to engage when tools feel human, conversational, and responsive.
What’s next for CivixAI
- Mobile-first deployment for broader community reach.
- Voice input and multilingual TTS for better accessibility.
- Integration with 911 and local emergency systems.
- Launching community trend dashboards for city planners and public health officials.
- Building an open-source framework so other cities can adopt CivixAI for their needs.
Log in or sign up for Devpost to join the conversation.