Skip to content

thereal-baitjet/SentinalAiAudit

Repository files navigation

Sentinal AI Audit

React TypeScript Tailwind CSS Gemini API AI Model License Status Platform Created By

Sentinal AI Audit is an advanced security intelligence tool designed to automate the analysis of CCTV and surveillance footage. By leveraging the multimodal capabilities of Google's Gemini 2.5 Flash model, it transforms raw video data into structured, searchable security logs.

🚀 Features

  • Automated Video Analysis: Upload CCTV footage (supports large files >20MB via Google File API) to detect entities, behaviors, and anomalies automatically.
  • Smart Event Timeline: Generates a precise, clickable timeline of events (Routine, Suspicious, Critical) allowing operators to jump instantly to key moments.
  • Executive Summaries: specific lighting conditions, duration analysis, and a high-level summary of the entire footage.
  • Security Grading: Automatically scores events based on severity (e.g., weapons or force detection vs. routine traffic).
  • Secure Architecture: API keys are handled securely via environment variables or dynamic selection, ensuring safe deployment.

🛠️ Tech Stack

  • Frontend: React 19, TypeScript
  • Styling: Tailwind CSS
  • AI/ML: Google Gemini API (@google/genai)
  • Model: gemini-2.5-flash

🔒 Security & Deployment

To deploy this application safely without exposing your personal API quota:

  1. Do NOT set the API_KEY environment variable in your hosting provider (Vercel, Netlify, etc.).
  2. Deploy the application as a static site.
  3. When you (or a user) opens the deployed site, click the Key Icon in the top right.
  4. Enter a Gemini API Key manually. This key is stored in the User's Browser Local Storage only.
  5. This ensures that every user provides their own key, and your personal key remains private and safe.

📦 How it Works

  1. Upload: The user uploads a video file. Small files are processed inline; large files are securely uploaded via the Gemini File API.
  2. Analyze: The Gemini 2.5 Flash model processes the visual data to identify security-relevant events, filtering out environmental noise (rain, trees).
  3. Report: The app renders a dashboard with a video player synced to a structured event log and intelligence summary.

👨‍💻 Credits

Created by Vibe.

This project was built to demonstrate the power of multimodal AI in physical security and surveillance auditing.

About

An Ai CCTV Audit Tech

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published