Skip to content

GEMINI.md

Latest commit

 

History

History
72 lines (58 loc) · 3.65 KB

GEMINI.md

File metadata and controls

72 lines (58 loc) · 3.65 KB

Project Overview

The BSOD AI Analyzer is an enterprise-grade application designed to diagnose Windows crash dump errors using Google's Gemini AI. It features a secure client-server architecture, with a React frontend for user interaction and an Express.js backend that proxies sensitive Gemini API calls. The application supports various crash dump formats, provides detailed analysis reports, and integrates advanced debugging capabilities. A key aspect of its design is robust security, including server-side API key protection and in-memory processing of crash dumps to ensure no sensitive data is stored.

Building and Running

This project uses Node.js and npm for dependency management and script execution.

Prerequisites:

  • Node.js 18+
  • Google Cloud account (for deployment)
  • Gemini API key from Google AI Studio

Local Development:

  1. Clone the repository:

    git clone https://github.com/faratech/bsod-analyzer.git
cd bsod-analyzer

  2. Install dependencies:

    npm install

  3. Set up environment: Create a .env.local file in the project root with your Gemini API key:

    GEMINI_API_KEY=your-gemini-api-key

    For additional secrets like TURNSTILE_SECRET_KEY and SESSION_SECRET, you can add them to .env.local or set them as environment variables directly.

  4. Start development server:

    npm run dev

    This command concurrently starts both the backend (on port 8080) and the frontend development server.

Development Commands:

  • npm run dev: Starts both backend and frontend servers.
  • npm run dev:backend: Starts the backend server only.
  • npm run dev:frontend: Starts the frontend development server only.
  • npm run build: Builds the production-ready frontend.
  • npm start: Runs the production server.

Deployment:

The application can be deployed to Google Cloud Run. The README.md provides detailed instructions for quick deployment using deploy-with-secret.sh script, manual deployment via Docker and gcloud CLI, and CI/CD setup with Cloud Build.

Development Conventions

  • Technology Stack:
    • Frontend: React 19, TypeScript, Vite
    • Backend: Express.js with ES modules
    • AI Service: Google Gemini 3.1 Flash Lite with grounding via @google/genai SDK
    • Styling: Custom CSS
    • Deployment: Docker, Google Cloud Run, Secret Manager
  • Secret Management:
    • Production: All secrets (e.g., GEMINI_API_KEY, TURNSTILE_SECRET_KEY, SESSION_SECRET) are stored in Google Secret Manager and injected as environment variables by Cloud Run.
    • Local Development: Secrets are managed via .env.local files or direct environment variables. The .env.local file is git-ignored and should never be committed.
    • Dedicated scripts (setup-all-secrets.sh, update-turnstile-secret.sh, deploy-with-secret.sh) are provided for secret management.
  • Security Best Practices:
    • Never commit secrets to the repository.
    • Regularly rotate secrets.
    • Utilize Secret Manager for production environments.
    • Adhere to the principle of least privilege for secret access.
  • Code Structure:
    • Frontend components are located in the components/ and pages/ directories.
    • Backend logic is primarily in server.js and services/.
    • Utility functions are in the utils/ directory.
  • Testing: (No explicit testing commands or frameworks were identified in the README.md beyond general development commands. Further exploration would be needed to determine specific testing conventions.)