AI Math Agent

An AI-powered mathematical assistant built with LangChain and Google Gemini.

---

Description

AI Math Agent is an intelligent chatbot that uses Google Gemini and LangChain to solve mathematical expressions accurately.

Instead of relying on the LLM's own calculations, the agent uses a custom calculator tool powered by SymPy for accurate mathematical operations.

The project also implements a custom conversation memory system that allows the AI to remember previous messages during the conversation.

---

Features

• 🤖 AI Agent powered by Google Gemini
• 🔧 Custom tool calling using LangChain
• 🧮 Accurate mathematical calculations using SymPy
• 🧠 Custom conversation memory system
• 💬 Maintains chat history during runtime
• 🏗️ Clean modular architecture
• 🔐 Secure API key management using environment variables

---

How It Works

User Input
     |
     v
Human Message
     |
     v
AI Agent (Gemini)
     |
     v
Tool Required?
     |
  +--+--+
  |     |
 No    Yes
  |     |
  v     v
Answer  Calculator Tool
             |
             v
        Tool Response
             |
             v
        Final AI Response

---

Project Structure

AI-Math-Agent/
│
├── agent.py              # Agent workflow and tool handling
├── tools.py              # Calculator tool implementation
├── memory.py             # Custom conversation memory
├── config.py             # Configuration management
├── main.py               # Application entry point
├── requirements.txt      # Project dependencies
├── .env.example          # Environment variable template
├── .gitignore            # Git ignored files
└── README.md             # Documentation

---

Installation

Clone Repository

git clone https://github.com/ridazahra-ai/AI-Math-Agent.git

Navigate to Project Folder

cd AI-Math-Agent

Create Virtual Environment

python -m venv venv

Activate Virtual Environment

Windows:

venv\Scripts\activate

Install Dependencies

pip install -r requirements.txt

---

Environment Setup

Create a .env file in the project directory.

Add your Google Gemini API key:

GOOGLE_API_KEY=your_api_key_here
MODEL_NAME=gemini-2.5-flash

---

Usage

Run the application:

python main.py

Example:

You: What is 25 * 17?

AI: The answer is 425.

Example with memory:

You: My name is Rida.

AI: Nice to meet you, Rida.

You: What is my name?

AI: Your name is Rida.

---

Technologies Used

• Python
• LangChain
• Google Gemini API
• LangChain Google GenAI
• SymPy
• python-dotenv

---

Concepts Implemented

• Large Language Model (LLM) Integration
• AI Agent Workflow
• Tool Calling
• Custom Conversation Memory
• Message-Based Conversation Handling
• Environment Configuration
• Modular Software Architecture

---

Future Improvements

• Implement LangGraph agent workflow
• Add persistent database memory
• Support multiple tools
• Improve error handling
• Build FastAPI backend
• Add frontend interface
• Deploy as a cloud-based AI service

---

License

MIT License