🌾 Tamil Farmer Assistant — AI Agent for Agriculture

A Tamil-language multi-tool AI assistant built for farmers in Jaffna, Sri Lanka. It combines live weather data, rule-based crop recommendations, and market price information with an LLM-powered conversational interface — so farmers can get practical advice entirely in Tamil.

Originally built as the capstone project for the Kaggle × Google 5-Day AI Agents Intensive (Vibe Coding) Course.

🚀 Live Demo

🔗 App: https://tamil-farmer-agent-j3cyzcehjeedshdjfmt3mw.streamlit.app/

Hosted on Streamlit Community Cloud, with UptimeRobot pinging every 5 minutes to prevent the free-tier app from sleeping.

✨ Features

• 🌤️ Live weather for Jaffna via the OpenWeatherMap API (temperature, humidity, conditions)
• 🌱 Crop suggestions generated from current temperature and humidity conditions
• 💰 Market price snapshot for common crops (rice, tomato, onion, chilli, banana, brinjal)
• 💬 Tamil-only conversational chat, powered by Groq's LLaMA 3.3 model, that stays grounded in the live weather/crop/price context for every answer
• 🧩 Two agent implementations included: a direct Groq chat-completion agent (agent.py) and a Google ADK multi-tool agent (adk_agent.py) that orchestrates the same three tools

🛠️ Tech Stack

• Language: Python
• UI: Streamlit
• LLM: Groq API — LLaMA 3.3 (70B)
• Agent Framework: Google Agent Development Kit (ADK) with Gemini 1.5 Flash
• External API: OpenWeatherMap
• Config: python-dotenv for environment variables

🏗️ Project Structure

Tamil-farmer-agent
├── app.py            # Streamlit UI — weather, crops, market price, chat
├── agent.py           # Groq-based conversational agent (CLI + chat logic)
├── adk_agent.py        # Google ADK multi-tool agent implementation
├── requirements.txt
└── tools/
    ├── weather.py       # OpenWeatherMap integration
    ├── crops.py         # Rule-based crop recommendation logic
    └── market.py        # Market price data

🏃 Getting Started

1. Clone the repository

git clone https://github.com/abilasha24/Tamil-farmer-agent.git
cd Tamil-farmer-agent

2. Install dependencies

pip install -r requirements.txt

3. Set environment variables

Create a .env file in the project root:

GROQ_API_KEY=your_groq_api_key
WEATHER_API_KEY=your_openweathermap_api_key

Never commit real API keys to GitHub.

4. Run the app

streamlit run app.py

📌 Project Highlights

• Multi-tool AI agent architecture combining external APIs, rule-based logic, and an LLM
• Fully Tamil-language system prompt and conversational responses
• Two parallel agent implementations (direct API vs. Google ADK) for comparison
• Deployed and monitored for uptime on a free-tier cloud host

👩‍💻 Author

Abilasha Selvanayakam

• GitHub: https://github.com/abilasha24
• LinkedIn: https://www.linkedin.com/in/abilashaselvanayakam2k06/
• Portfolio: https://my-portfolio-webapp-ashy.vercel.app/