🧠 TinyAGI

TinyAGI is a powerful, modular Artificial General Intelligence (AGI) framework crafted for seamless integration and management of AI agents, plugins, and tools. With its adaptable and extensible architecture, TinyAGI enables dynamic loading of components from local and GitHub-hosted sources, empowering you to customize and scale for a multitude of use cases. It's designed to be lightweight, easy to understand, and highly extensible.

Disclaimer: This is a preview release (version 0.0.2) of TinyAGI, which is currently under active development. This release is intended for testing, feedback, and early experimentation. Please be aware that the API is subject to change, and frequent updates are expected.

---

📖 Table of Contents

• 🧩 Features
• 📦 Installation
• 🔧 Setup Instructions
• 🛠️ Usage
  • Interactive CLI
  • Direct Commands
  • API Server
  • Programmatic Usage
• 📚 Documentation
• 🧪 Testing
• 📝 Contributing
• 🛡 License
• 📞 Contact
• 🗺️ Roadmap

---

🧩 Features

• 🌐 Agent System: Manage various AI agents, from OpenAI to Ollama and LLaMA.
• 🔌 Plugin Manager: Seamlessly expand functionality with versatile plugins.
• 🛠 Tool Integration: Use tools like the Wikipedia API to enhance capabilities.
• 🔄 Dynamic Loading: Load components locally or clone from GitHub repositories on the fly.
• 🚀 Task Automation: Orchestrate agents, plugins, and tools to define and execute complex tasks.
• 🖥️ Dual Interface: Interact via a feature-rich interactive CLI or a RESTful API server.
• 📑 Comprehensive Documentation: Easily accessible Markdown files for every component.
• 🔥 Robust Error Handling: Advanced logging and error management for smooth operation.

---

📦 Installation

The recommended way to install TinyAGI is by using the installation scripts, which automate the entire process from downloading the code to setting up the environment. The installation script will handle everything:

• Check for prerequisites (Git, Python).
• Clone the project to a fixed location (C:\TinyAGI on Windows).
• Create a dedicated virtual environment.
• Install all required dependencies.
• Create a default .env file for your API keys.

1. Clone the Repository

First, clone the repository to your local machine:

git clone https://github.com/SullyGreene/TinyAGI.git
cd TinyAGI

2. Run the Installation Script

<details> <summary><b>For Windows</b></summary>

Run the install.bat script. It will check for prerequisites (Git, Python), set up a virtual environment, install dependencies, and create a default .env file.

install.bat

The script will install TinyAGI to C:\TinyAGI by default. You can customize this path by creating a config\windows_installation.ini file in the cloned directory before running the script, and setting custom_install_path=Your\Desired\Path. The script will also offer to add an agi command alias to your system PATH for easy access to the CLI. Note: Run this script from a temporary location after cloning, as it will move the project files to the designated installation directory.

</details> <details> <summary><b>For macOS / Linux</b></summary>

Run the shell script to set up the environment and install dependencies.

chmod +x install.sh
./install.sh

</details>

3. Configure Environment Variables

After the installation script completes, a .env file will be created in the root directory. Open this file and add your API keys (e.g., for OpenAI, Gemini) to enable the corresponding agents.

# .env
OPENAI_API_KEY="your-key-here"
GEMINI_API_KEY="your-key-here"

---

📖 Programmatic Usage

You can easily integrate TinyAGI into your own Python projects. The core of the framework is the AgentSystem, which orchestrates agents, plugins, and tools to execute tasks defined in a configuration file.

Here is a basic example of how to run the agent system from a Python script:

import TinyAGI as agi
import os

def main():
    # Ensure the config path is correct relative to your script
    config_path = os.path.join(os.path.dirname(__file__), 'config', 'agent_config.json')
    
    # Initialize and run the agent system
    agent_system = agi.AgentSystem(config_files=config_path)
    agent_system.run()

if __name__ == '__main__':
    main()

🛠 Usage

Running the Interactive CLI

<details> <summary><b>For Windows</b></summary>

cli.bat

</details> <details> <summary><b>For macOS / Linux</b></summary>

poetry run cli

</details>

This will start the interactive CLI, where you can use the following commands:

• generate: Generate text from a prompt.
• config: Display current configuration.
• exit: Exit the CLI.

Starting the Server

<details> <summary><b>For Windows</b></summary>

start_server.bat

</details> <details> <summary><b>For macOS / Linux</b></summary>

poetry run start

</details>

Accessing the API

• Chat Endpoint

  <details> <summary>Show command</summary>

  curl -X POST http://localhost:5000/chat \
       -H "Content-Type: application/json" \
       -d '{"messages": [{"role": "user", "content": "Hello!"}], "stream": false}'
  

  </details>

• Generate Text Endpoint

  <details> <summary>Show command</summary>

  curl -X POST http://localhost:5000/generate \
       -H "Content-Type: application/json" \
       -d '{"prompt": "Write a short story about a dragon.", "stream": false}'
  

  </details>

• Embed Endpoint

  <details> <summary>Show command</summary>

  curl -X POST http://localhost:5000/embed \
       -H "Content-Type: application/json" \
       -d '{"input": "Sample text for embedding."}'
  

  </details>

---

📚 Documentation

Access comprehensive documentation in the README.md files in each directory.

---

🧪 Testing

Run TinyAGI tests to verify functionality.

Testing Ollama

<details> <summary>Show command</summary>

python test_ollama_agent.py

• Expected Output:

  Response from OllamaAgent:
  The capital of France is Paris.
  

• Troubleshooting Tips:

  • Ensure the Ollama server is running at http://localhost:11434.
  • Confirm correct API keys and authentication.

</details>

---

🧪 Testing

Run the test suite to ensure all components are working correctly. Make sure you have installed the development dependencies with poetry install.

<details> <summary>Show command</summary>

poetry run pytest

</details>

---

📝 Contributing

Join the TinyAGI community by contributing your code, ideas, or feedback!

1. Fork the Repository

   <details> <summary>Show commands</summary>

   git checkout -b feature/YourFeatureName
   

   </details>

2. Commit Your Changes

   <details> <summary>Show command</summary>

   git commit -m "Add feature: YourFeatureName"
   

   </details>

3. Push to Your Fork

   <details> <summary>Show command</summary>

   git push origin feature/YourFeatureName
   

   </details>

4. Submit a Pull Request

---

🛡 License

This project is licensed under the MIT License.

---

📞 Contact

💬 Get in Touch

• X (formerly Twitter): @SullyGreene
• GitHub: SullyGreene
• PyPI: TinyAGI on PyPI

---

🗺️ Roadmap

TinyAGI’s vision includes scaling to meet diverse AI needs. Here’s what’s planned:

<details> <summary>Phase 1: Core Enhancements 🚀</summary>

• Agent Expansion: Support additional agents and tools for specific domains.
• Plugin Ecosystem: Expand with plugins for data analysis, visual generation, and task-specific fine-tuning.
• Advanced Error Handling: Improve diagnostic logs and error handling.

</details> <details> <summary>Phase 2: Advanced Task Orchestration 🤖</summary>

• Multi-Agent Collaboration: Enable agents to collaborate on complex tasks.
• Task Scheduling & Automation: Automate recurring actions and analysis.
• Smart Prompting: Dynamic prompt optimization for better task performance.

</details> <details> <summary>Phase 3: Enhanced API and Developer Experience 🛠️</summary>

• API V2: Improve task queueing, agent behavior management, and access controls.
• Interactive Documentation: Launch an interactive portal with live code examples.
• CLI Improvements: Add user-friendly CLI commands.

</details> <details> <summary>Phase 4: Ecosystem & Community Growt