Inspiration
Millions of everyday farmers across Africa make critical decisions daily about what to plant, how to treat crop disease, and how to protect their harvest, without access to expert agricultural advice. Most AI tools only work in English, excluding the majority of African farmers who need them most.
I built FarmAssist to change that.
What it does
FarmAssist is an AI-powered farm advisor that gives everyday farmers instant, structured advice on:
- Crop disease identification and treatment
- Planting guidance based on season and location
- Pest detection tips
- Fertilizer recommendations
What makes FarmAssist different is multilingual support—it automatically detects the farmer's language and responds in it. English, Yoruba, Hausa, Igbo, French, and Swahili.
Farmers can also use voice input—speak directly to FarmAssist without typing.
How we built it
- Amazon Nova 2 Lite — core AI intelligence, farming advice and multilingual responses
- Next.js 14 — frontend framework
- Web Speech API — voice to text input
- Vercel — deployment
- Tailwind CSS — styling
Challenges we ran into
Prompting engineering Nova to consistently respond in the user's detected language while maintaining structured, practical farming advice was the biggest challenge. Getting the tone right for farmers—simple, clear, and actionable—required significant iteration.
Accomplishments that we're proud of
Successfully integrating Amazon Nova 2 Lite to deliver structured, practical farming advice instantly
- Building genuine multilingual support that automatically detects and responds in Yoruba, Hausa, Igbo, French, and Swahili.
- Designing a clean, accessible UI that a farmer with limited tech experience can navigate easily
- Adding voice input so farmers can speak directly without typing
- Building and shipping a complete product in under 48 hours
What we learned
Building for a specific underserved community forces you to think beyond features. Every design decision had to serve a farmer with limited tech experience, not a developer. Prompting Nova to maintain both language accuracy and a practical farming tone simultaneously was harder than expected—and more rewarding when it worked.
What's next for FarmAssist - AI Farm Advisor for African Farmers
- Image analysis — farmers upload a photo of their crop and Nova identifies the disease visually
- Nova Act integration — automatically find nearest agricultural supply stores based on farmer's location
- Offline mode — critical for rural farmers with limited connectivity
- Expanding African languages — adding more local languages across West and East Africa
- SMS integration — reaching farmers who don't have smartphones
Built With
- amazon-nova-lite-2
- nextjs
- tailwindcss
- typescript
- vercel
- web-speech-api
Log in or sign up for Devpost to join the conversation.