Make Your Own AI Agent

A comprehensive demo collection showing how to build AI agents using LangChain with different models and tools.

🚀 Quick Start

Prerequisites

1. Python 3.8+ installed
2. uv for fast Python package management:

   # Install uv
   curl -LsSf https://astral.sh/uv/install.sh | sh
   # or with pip
   pip install uv

3. Project setup:

   # Create virtual environment and install dependencies
   uv venv
   source .venv/bin/activate  # On Windows: .venv\Scripts\activate
   uv pip install langchain langchain-ollama langchain-google-genai

4. API Keys (for cloud models):
   • Google Gemini: Get API key from Google AI Studio
5. Ollama (for local models):
   • Install from ollama.ai
   • Pull a model: ollama pull gpt-oss:20b

📁 Demo Files

01_use_chat_model.py

Basic Chat Model Usage with Google Gemini

Learn how to:

• Initialize chat models with API keys
• Send basic chat messages
• Handle multi-turn conversations
• Use streaming responses
• Batch process multiple queries
• Implement async processing

uv run python 01_use_chat_model.py

02_use_tools.py

Tool Creation and Integration

Explore:

• Creating custom tools with @tool decorator
• Binding tools to chat models
• Manual tool execution
• Tool execution loops
• Advanced tools with complex return types
• Error handling in tools

uv run python 02_use_tools.py

03_use_ollama.py

Local AI with Ollama

Discover:

• Setting up ChatOllama for local inference
• Using different models (llama2, codellama, mistral, etc.)
• System prompts and conversation management
• Streaming responses
• Custom model parameters
• Model availability checking

uv run python 03_use_ollama.py

04_use_tools_with_ollama.py

Combining Local AI with Tools

Master:

• Integrating tools with Ollama models
• File operations (read/write)
• Multi-step tool execution chains
• Weather and calculation combos
• Error handling with tools
• Context-aware tool usage

uv run python 04_use_tools_with_ollama.py

🛠️ Setup Instructions

For Google Gemini (Cloud)

1. Ensure virtual environment is activated:

   source .venv/bin/activate  # On Windows: .venv\Scripts\activate

2. Get your API key from Google AI Studio

3. Set environment variable:

   macOS/Linux:

   export GOOGLE_API_KEY="your-api-key-here"
   # To make it permanent, add to ~/.bashrc or ~/.zshrc:
   echo 'export GOOGLE_API_KEY="your-api-key-here"' >> ~/.bashrc
   # or for zsh:
   echo 'export GOOGLE_API_KEY="your-api-key-here"' >> ~/.zshrc

   Windows (Command Prompt):

   set GOOGLE_API_KEY=your-api-key-here
   # To make it permanent:
   setx GOOGLE_API_KEY "your-api-key-here"

   Windows (PowerShell):

   $env:GOOGLE_API_KEY="your-api-key-here"
   # To make it permanent:
   [System.Environment]::SetEnvironmentVariable('GOOGLE_API_KEY', 'your-api-key-here', 'User')

4. Run the demos:

   uv run python 01_use_chat_model.py
   uv run python 02_use_tools.py

For Ollama (Local)

1. Install Ollama:

   # macOS
   brew install ollama
   
   # Linux
   curl -fsSL https://ollama.ai/install.sh | sh

2. Start Ollama server:

   ollama serve

3. Pull a model:

   ollama pull gpt-oss:20b
   # or try other models:
   # ollama pull llama2
   # ollama pull codellama
   # ollama pull mistral

4. Run local demos:

   uv run python 03_use_ollama.py
   uv run python 04_use_tools_with_ollama.py

🎯 Use Cases

Cloud AI (Google Gemini)

• Production applications requiring high accuracy
• Applications with internet connectivity
• Complex reasoning tasks
• Multilingual support

Local AI (Ollama)

• Privacy-sensitive applications
• Offline environments
• Cost-effective solutions
• Development and experimentation

🔧 Available Tools

The demos include several pre-built tools:

• Calculator: Safe mathematical expression evaluation
• File Reader/Writer: Text file operations
• Current Time: Date and time information
• Weather: Mock weather data (extensible to real APIs)
• Word Counter: Text analysis and statistics
• Number Analyzer: Statistical analysis of number arrays
• File Search: Pattern-based file searching

🚦 Running the Demos

Basic Usage

# Activate virtual environment first
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

# Cloud-based AI with tools
uv run python 01_use_chat_model.py
uv run python 02_use_tools.py

# Local AI with Ollama
uv run python 03_use_ollama.py
uv run python 04_use_tools_with_ollama.py

Modifying Examples

Each demo file is heavily commented and modular. You can:

1. Uncomment sections to enable/disable specific examples
2. Modify prompts to test different scenarios
3. Add new tools by following the @tool decorator pattern
4. Change models by updating the model names
5. Adjust parameters like temperature, top_p, etc.

🔍 Troubleshooting

Common Issues

Google Gemini API Errors:

Error: API key not found

• Set your GOOGLE_API_KEY environment variable:
  • macOS/Linux: export GOOGLE_API_KEY="your-key"
  • Windows CMD: setx GOOGLE_API_KEY "your-key"
  • Windows PowerShell: $env:GOOGLE_API_KEY="your-key"
• Verify the key is valid in Google AI Studio
• Restart your terminal/IDE after setting permanent environment variables

Ollama Connection Errors:

Error connecting to Ollama

• Check if Ollama is running: ollama ps
• Verify the model is installed: ollama list
• Start Ollama server: ollama serve

Import Errors:

ModuleNotFoundError: No module named 'langchain_ollama'

• Install missing dependencies: uv pip install langchain-ollama
• Or activate your virtual environment: source .venv/bin/activate

📚 Learning Path

1. Start with 01_use_chat_model.py to understand basic chat interactions
2. Move to 02_use_tools.py to learn tool creation and binding
3. Try 03_use_ollama.py for local AI model usage
4. Complete with 04_use_tools_with_ollama.py for advanced local AI + tools

🤝 Contributing

Feel free to:

• Add new demo files
• Create additional tools
• Improve error handling
• Add support for other models
• Update documentation

📄 License

This project is open source and available under the MIT License.

🆘 Support

• LangChain Documentation: python.langchain.com
• Ollama Documentation: ollama.ai
• Google Gemini: ai.google.dev

---

Happy AI Agent Building! 🤖