An agent that acts as a bridge between visual reality (video/images) and technical documentation (PDFs/Manuals) to detect contradictions and audit procedures.
- Node.js 18+
- npm or yarn
- Gemini API Key (get one at https://aistudio.google.com/app/apikey)
-
Install dependencies:
npm install
-
Configure API Key:
Create a
.envfile in the project root:cp .env.example .env
Then edit
.envand add your API key:API_KEY=your_gemini_api_key_hereOr export as environment variable:
export API_KEY=your_gemini_api_key_here
Start the development server:
npm startThe app will be available at http://localhost:4200
Build for production:
npm run buildThe built files will be in dist/protocol-eye/browser/
- Provide SOP: Paste the Standard Operating Procedure (SOP) text directly or upload a text file (.txt, .md) in the left panel.
- Upload Evidence: Upload an image or video frame showing the procedure being performed.
- Audit: Click "Run Audit" to analyze the visual evidence against the SOP.
- Review Results: Examine the audit results including:
- Executive Summary (Pass/Fail/Caution)
- Protocol Matches
- Discrepancy Log
- Reasoning Trace
├── src/
│ ├── app.component.ts # Main Angular component
│ ├── app.component.html # Template
│ ├── environments/ # Environment configs
│ └── models/ # TypeScript interfaces
├── services/
│ └── gemini.service.ts # Gemini API integration
├── index.tsx # Application bootstrap
├── index.html # HTML entry point
├── angular.json # Angular CLI config
├── tsconfig.json # TypeScript config
└── package.json # Dependencies
- Angular 21
- Google Gemini 2.5 Flash
- Tailwind CSS
- TypeScript