FarmGPT AI splits farm tasks across specialized AI agents
FarmGPT AI is an AI farming assistant built to support decisions across the crop lifecycle. It can detect crop diseases from leaf images, answer farming questions, create farm plans as downloadable PDF reports, analyze crop prices, and provide . It also includes sign-in, farmer profiles, and a mobile-friendly interface.
Instead of sending every task through one large prompt, an directs each request to a specialized agent for disease detection, farm planning, market , or overall recommendations. This structure made the easier to organize and expand. Designing the workflow and routing requests between agents were the hardest parts of the build.
The project combines React, , Supabase PostgreSQL, Gemini AI, Vercel, and other web tools.
Key points
- An sends each request to the appropriate specialized agent.
- Separate agents handle disease detection, farm planning, market data, and overall recommendations.
- The product includes leaf-image checks, chat, PDF plans, price analysis, and personalized advice.
- Workflow design and request routing were the most difficult parts to build.
- No token or cost comparison was provided.