Turing AI Agent

A locally running AI agent that can use tools, search the web, and remember information across sessions.

Python · Ollama · Qwen · SQLite · Tavily

🧠 What is Turing?

Turing is a local AI agent built with Python and Ollama.

Instead of only generating text, Turing can use external tools, retrieve real-world information, search the web, and store information in persistent memory.

Current capabilities

• 🧮 Calculator
• 🌤️ Current weather
• ⏰ Time and time zones
• 🌐 Web search
• 🧠 Persistent memory
• 🔗 Multiple tool calls
• 📝 Runtime logging
• 🛡️ Error handling

⚙️ How It Works

User → Turing → Local LLM → Tool Registry → Tools → Results → Local LLM → Response

Turing keeps its tools separate from the main agent through a tool registry, making the system easier to extend.

🔧 Tools

🧠 Persistent Memory

Turing can store information in a local SQLite database and retrieve it after the program is restarted.

User → Memory Tool → SQLite → Future Conversation → Turing retrieves the information

📁 Project Structure

turing-ai-agent/
│
├── main.py
├── README.md
├── requirements.txt
├── .env
├── .gitignore
│
├── data/
│   └── memory.db
│
├── logs/
│   └── agent.log
│
└── tools/
    ├── calculator.py
    ├── weather.py
    ├── time.py
    ├── memory_tool.py
    ├── web_search.py
    └── registry.py

.env contains private API credentials and should never be committed to GitHub.

🛠️ Tech Stack

• Python — Agent logic and tool execution
• Ollama — Local LLM runtime
• Qwen 2.5 — Language model
• SQLite — Persistent memory
• Tavily — Web search
• Open-Meteo — Weather data
• python-dotenv — Environment configuration

🚀 Run Locally

1. Clone the repository

git clone <repository-url>
cd turing-ai-agent

2. Install dependencies

pip install -r requirements.txt

3. Download the model

ollama pull qwen2.5:1.5b

Make sure Ollama is running.

4. Configure Tavily

Create a .env file:

TAVILY_API_KEY=your_api_key_here

5. Run Turing

python main.py

🗺️ Roadmap

Completed

• Local LLM integration
• Conversational context
• Tool calling
• Dynamic tool registry
• Calculator
• Weather
• Time
• Persistent SQLite memory
• Web search
• Multiple tool calls
• Logging
• Error handling

Next

• Improved agent loop
• Multi-step autonomous tasks
• Tool validation
• Conversation persistence
• Relevant memory retrieval
• Async tool execution
• Automated testing
• FastAPI backend
• Local web interface
• Document processing
• RAG

🎯 Why I'm Building Turing

Turing is a hands-on project for understanding how AI agents are actually engineered — combining language models with tools, APIs, databases, memory, and eventually autonomous multi-step workflows.

The goal isn't simply to build another chatbot.

It's to build an agent whose capabilities and architecture can grow over time.

👨‍💻 Author

Zermello

Built with Python, curiosity, and a lot of debugging. 🤖

🚧 Turing is an evolving project. New capabilities are being added as the architecture develops.