Inspiration
The project was inspired by the need to bridge technology and legal empowerment for harassment victims. Many people feel powerless and unsure of their rights when incidents occur. We wanted to create an accessible tool that combines AI analysis with comprehensive NSW law guidance, making justice more accessible to those who need it most.
What it does
The platform combines voice recording, AI-powered harassment analysis, and NSW legal guidance. Users can record incidents live, upload audio files, or type descriptions. The AI analyzes content for harassment patterns and severity, then provides specific legal information including relevant laws, punishment ranges, and recommended actions based on case severity.
How we built it
Frontend: Next.js 14 with TypeScript and Tailwind CSS for a professional, government-like interface.
AI Integration: Ollama (llama2) for harassment analysis with robust fallback systems.
Audio Processing: Web Speech API for live transcription, OpenAI Whisper for file uploads.
Legal Framework: Comprehensive NSW law database with dynamic mapping to harassment types.
Architecture: Component-based structure with responsive design optimized for mobile use.
Challenges we ran into
AI Reliability: Ollama model initially refused transcription requests due to ethical concerns, requiring fallback systems.
Audio Processing: Web Speech API limitations with uploaded files led to OpenAI Whisper integration.
Legal Accuracy: Ensuring accurate NSW law information and appropriate legal disclaimers.
Technical Issues: Hydration errors, TypeScript integration, and performance optimization challenges.
User Experience: Designing for stressful situations while maintaining professional credibility.
Accomplishments that we're proud of
Seamless AI Integration: Successfully integrated multiple AI services with robust fallbacks Professional UI/UX: Created a government-like interface that builds user trust Comprehensive Legal Coverage: Built system covering all major NSW harassment laws Real-time Processing: Achieved near-instant harassment analysis and legal guidance Mobile-First Design: Optimized for smartphone use during incident documentation
What we learned
AI Limitations: Understanding when and how to use different AI services effectively.
Legal Technology: The complexity of presenting legal information in digestible formats.
User Experience: How to design interfaces for users in crisis situations.
Technical Development: Advanced Next.js 14 features and TypeScript best practices.
Fallback Systems: The importance of robust backup systems when primary services fail.
What's next for NSW Anti Harassment AI Platform
Immediate: Multi-language support, enhanced AI models, mobile apps, offline capability.
Advanced Features: Community incident mapping, legal document generation, police integration, AI mental health support.
AI Improvements: Better context understanding, risk assessment, multi-modal analysis.
Legal Expansion: Coverage for other Australian states, international laws, case law integration.
Community Features: Support groups, educational resources, advocacy tools, research platform.
Built With
- css
- javascript
- llma2
- nextjs
- openai-api
- whisperai


Log in or sign up for Devpost to join the conversation.