J-Jaga

J-Jaga, an AI-powered accident guardian designed for the most STRESSFUL moment a driver can face, which is the minutes right after a crash.

Comment

Inspiration

Project Story: J-Jaga, Our Guardian When It Matters Most

J-Jaga (meaning: to guard or watch over in Malay) was inspired by a personal experience. Shortly after my friend and I got our driving licenses, we were involved in a car accident. While the collision itself was not severe, the situation quickly became overwhelming. We broke down, couldn’t reach anyone, and the car didn’t have a dashcam to capture what happened.

From this experience, I learned that most drivers are unprepared for the critical minutes after an accident. Panic leads to missed evidence, poor communication, and long-term consequences with insurance or legal claims.

J-Jaga leverages Google's Gemini 3 API as its core intelligence layer, utilising multiple advanced features to create an automotive AI assistant. The application employs multimodal processing to simultaneously analyse live camera feeds, engine audio recordings, and natural language conversations, enabling real-time incident documentation and vehicle diagnostics. The Guardian mode uses Gemini's Live API with function calling to process streaming video during accidents, intelligently deploying three custom tools: draw_ar_marker for AR-guided evidence capture, holographic_overlay for HUD data panels, and log_evidence for timestamped incident recording. System instructions guide the AI to prioritise injury detection and witness statements with Malaysian traffic context awareness.

At its core: $$Calm Guidance +Proper Documentation =Better Protection$$

The biggest challenge was designing an experience that feels reassuring rather than overwhelming. This project taught us that technology is most powerful when it supports people at their most vulnerable.

Built With

Share this project:

Updates