Access to concise, actionable farming information is difficult, especially in resource-limited settings or areas with intermittent internet connectivity.

Farmers need readily available guidance on crop cultivation and quick identification of potential plant health issues to maximize yield and prevent losses.

AgriSolver is a lightweight web application designed to provide farmers with essential agricultural information. It combines AI-powered summarization for grow/plant instructions, offering offline capabilities for core information retrieval.

1. Plant Instruction Summarization:

   • Input a plant name (e.g., "tomato", "eggplant").
   • Leverages the Perplexity AI API 0 to generate concise growing/planting instructions.
   • Caches search results in a local SQLite database, allowing access to previously searched information even without an internet connection.

2. Australian good agriculture example:

   • Use Australia blueberry 40 year history to inspire farming decisions.

• Frontend: React (Vite)
• Backend: Node.js, Express.js
• Database: SQLite
• AI Services:
  • Perplexity AI API (for summarization)

• Node.js and npm installed
• API Keys for Perplexity AI and Google Gemini (Store in backend/.env)

1. Navigate to the backend directory: cd backend
2. Install dependencies: npm install (this will now include axios)
3. Create a .env file in the backend directory and add your API keys:

   PERPLEXITY_API_KEY=your_perplexity_api_key

4. Start the backend server for development: npm run dev. This will automatically create the agrisolver.db file if it doesn't exist.
5. Start the backend server for production: npm start

1. Navigate to the frontend directory: cd frontend
2. Install dependencies: npm install
3. Start the frontend development server: npm run dev
4. Open your browser and navigate to the URL provided by Vite (usually http://localhost:5173).

Ensure you have obtained API keys from Perplexity AI and Google Cloud (for Gemini API). Store these securely in the .env file in the backend directory. Do not commit the .env file to version control.

• Deployment: Containerize using Docker for easy deployment.
• Scalability: Evaluate database alternatives for larger scale (e.g., PostgreSQL).
• Offline Enhancements: Explore ways to make the image recognition model available offline (potentially using smaller, on-device models if feasible).
• User Management: Add user accounts to save preferences and plans.