Local AI Chat Interface

Simple Python integration with Ollama for AI task execution, now with Strand Agent for enhanced chat handling. Includes built-in security guardrails for prompt injection prevention and input sanitization.

Prerequisites

• Python 3.8 or higher
• Ollama installed and running locally
• Internet connection for package installation

NOTE: Internet connection is only required for initial setup (installing dependencies and downloading models). During execution, the system runs entirely offline as both Strand Agent and Ollama operate locally.

Setup

1. Install Ollama and pull a model:

ollama pull llama3.2

2. Install dependencies:

pip install -r requirements.txt

Usage

Interactive Mode

python3 interactive_agent.py

Strand Agent CLI

# Chat with Strand Agent handling
python3 strand_cli.py chat "What is machine learning?"

# Streaming with Strand Agent
python3 strand_cli.py stream "Tell me a story"

Python API

Strand Agent Integration

from strand_ollama_agent import StrandOllamaAgent

agent = StrandOllamaAgent()

# Chat handled by Strand Agent
response = agent.chat("Hello, how are you?")
print(response)

# Streaming with callback
def print_chunk(chunk):
    print(chunk, end='', flush=True)

agent.stream_chat("Tell me a story", callback=print_chunk)

Original Agent

from ollama_agent import OllamaAgent

agent = OllamaAgent()

# Simple chat
response = agent.chat("Hello, how are you?")
print(response)

# Streaming chat with callback
agent.stream("Tell me a story", callback=print_chunk)

# Structured task execution
result = agent.task("analyze_data", {
    "numbers": [1, 2, 3, 4, 5],
    "operation": "calculate_statistics"
})
print(result)

Configuration

• Default URL: http://localhost:11434
• Default Model: llama3.2:latest
• Customizable: Pass different base_url and model to constructor

Project Structure

├── ollama_agent.py         # Core agent class
├── strand_ollama_agent.py  # Strand Agent integration
├── interactive_agent.py    # Interactive CLI
├── cli.py                 # Command line interface
├── strand_cli.py          # Strand Agent CLI
├── requirements.txt       # Dependencies
└── README.md             # This file