Inspiration
Every day, millions of farmers make decisions that affect food production, profitability, sustainability, and food security.
The knowledge required to make those decisions already exists.
But it is fragmented across scientific papers, government publications, agricultural websites, forums, PDFs, and disconnected databases.
Modern AI models are incredibly powerful.
But they can only reason with the knowledge they receive.
We realized the real challenge wasn't building another AI assistant.
It was building the knowledge AI could actually trust.
That idea became AgroDex.
What it does
AgroDex is an AI-powered agricultural knowledge platform.
Instead of treating agricultural information as isolated articles, AgroDex organizes it into structured relationships.
Every crop becomes connected to:
- Diseases
- Pests
- Symptoms
- Cultivation practices
- Climate
- Scientific references
- Images
- Related agricultural entities
This allows AI to navigate agricultural knowledge instead of simply generating generic answers.
For OpenAI Build Week we built Agro Smart, a dedicated demonstration showing how AI can reason over structured agricultural knowledge.
The demo includes:
- AI-assisted tomato diagnosis
- Potato ↔ Late Blight relationship exploration
- Aphid host relationship discovery
- Interactive agricultural knowledge graph
- Navigation through connected agricultural entities
The goal is not to replace agronomists.
The goal is to help people access trustworthy agricultural knowledge more naturally.
How we built it
AgroDex combines modern web technologies with OpenAI models to organize and explore structured agricultural knowledge.
For OpenAI Build Week we created a dedicated interactive demonstration based on real agricultural relationships from our knowledge base.
Instead of relying on artificial demo data, Agro Smart uses structured relationships between crops, diseases, pests, symptoms, and scientific references.
We also created a keynote-style product trailer to communicate the vision behind the project.
Challenges we ran into
The biggest challenge wasn't writing code.
The biggest challenge was organizing agricultural knowledge into relationships that AI can actually reason over.
Building a trustworthy knowledge layer required significantly more work than building the interface itself.
Another challenge was creating a demonstration that remained deterministic and reliable without depending on external AI services during the presentation.
Accomplishments that we're proud of
- Built a structured agricultural knowledge platform.
- Created Agro Smart, a dedicated AI demonstration for OpenAI Build Week.
- Connected crops, diseases, pests, symptoms, and scientific references into navigable relationships.
- Produced a keynote-style product trailer.
- Built a stable demonstration designed specifically for live presentations.
What we learned
Modern AI dramatically lowers the barrier for turning ambitious ideas into working software.
I didn't come from a traditional software engineering background.
Development started on June 9, 2026.
In just a few weeks, AgroDex evolved from an idea into a working AI-powered agricultural knowledge platform.
What started almost as a joke quickly became an obsession.
Modern AI tools made it possible to transform an idea into a working product much faster than would have been possible before.
The experience completely changed how I think about software development.
AI didn't replace my work.
It amplified my ability to build something I couldn't have built alone.
What's next for AgroDex
AgroDex is only the beginning.
Our long-term vision is an ecosystem built on three layers:
AgroDex — the knowledge layer.
Agro Smart — the reasoning layer.
AgroFarm — an offline-first farm management application powered by AgroDex intelligence.
We believe the future of agricultural AI starts with trusted knowledge.
Not bigger prompts.
Not bigger models.
Better knowledge.
Because...
AI is only as good as the knowledge behind it.
That's why we built AgroDex.
Built With
- codex
- css
- gpt-5.6
- next.js
- node.js
- openai
- postgresql
- react
- tailwind
- typescript
Log in or sign up for Devpost to join the conversation.