Gemination

Inspiration

This project was inspired by the real need of small farmers to get fast and simple crop diagnosis. Many farmers do not have easy access to experts, and crop diseases can spread very quickly. We also wanted to make the experience easier by supporting multiple languages and voice, so farmers can use the app in the way that feels most natural to them.

What it does

Gemination helps farmers diagnose crop problems using an image or video of their plant. The system analyzes the crop and returns:

• the crop type
  
• the possible disease
  
• the recommended treatment
  
• a confidence score
  

The app also supports many common languages. Farmers can chat with the AI in their own language and listen to responses using voice, which makes the app usable even for people who are not comfortable reading long text.

How we built it

We built Gemination using two main frameworks:

• FastAPI for the backend
  
• Next.js for the frontend
  

The backend uses Gemini AI as the main brain of the system. Gemini is used to analyze images and videos, understand and translate text, and transcribe audio messages. We also used Google Text-to-Speech to convert text responses into spoken audio.

For data storage, we used SQLAlchemy as our database layer.

Challenges we ran into

One of the biggest challenges was that this was our first large project. It was difficult at first to understand how different parts of the system should be organized and how they should communicate with each other.

Another challenge was working with the Gemini API for the first time. Even though the documentation was helpful, it still required time to understand how to adapt it to a real application.

We also faced time limitations, so we focused on building the core features that make the app useful and functional.

Accomplishments that we're proud of

We are very proud of turning an idea into a working product. Building our own crop diagnosis model was not realistic for us due to time and knowledge limits. Using Gemini allowed us to focus on creating useful features and a smooth user experience.

We are proud that the app is simple to use, practical, and designed to support small-scale farmers in real situations.

What we learned

We learned how to build a complete application, starting from an idea and ending with a working system. We also learned how to properly use the Gemini API and integrate its powerful features into a real project.

Most importantly, we learned how different parts of a system — AI, backend, frontend, database, and voice can work together as one product.

What's next for Gemination

Next, we plan to add a community feature where farmers can ask questions and experts can provide help, especially when the AI’s confidence is low.

We also want to build a marketplace where farmers can buy and sell farm products and treatments directly within the app.

Finally, we plan to expand language support, especially for African languages, to make the platform more accessible and useful for farmers across Africa.

Built With

• fastapi
• gemini
• google-web-speech-api
• nextjs
• pydantic
• python
• sqlalchemy