Skip to main content

Introduction

The AGI Framework: The Open-Source Framework for Artificial General Intelligence: Modular, Model-Agnostic, Scalable, and Democratized AI Development.

Version 0.6.4​

This is the introduction to the AGI Framework. If you are looking for more detailed project documentation, please read The complete AGI Framework Documentation. If you are looking for a high-level project overview, you're in the right place.

Why the AGI Framework Matters​

In today's rapidly evolving technological landscape, access to advanced AI capabilities should be universal and equitable. The AGI Framework represents a commitment to democratizing artificial general intelligence through:

  • Universal Access: Open-source, MIT-licensed architecture enabling unrestricted use and modification
  • Technological Equity: Ensuring AGI benefits are accessible to all segments of society
  • Innovation Potential: Fostering new opportunities across all economic sectors
  • Global Collaboration: Encouraging international cooperation in AGI development
  • Sustainable Development: Promoting responsible advancement that supports social stability

The AGI Framework isn't just a technological tool - it's a catalyst for positive societal transformation. By making AGI technology freely available, we aim to:

  • Enable new economic opportunities and business models
  • Support the evolution of more equitable economic frameworks
  • Foster innovation through unrestricted access to advanced AI capabilities
  • Promote global collaboration in addressing technological challenges

Introduction​

The AGI Framework is a pioneering open-source architecture that transcends the limitations of current AI systems, including Large Language Models (LLMs). It provides a modular and model-agnostic infrastructure that seamlessly integrates diverse AI models and tools, enabling intent-driven actions and autonomous execution. With a strong emphasis on scalability, adaptability, and ethical alignment, the framework is designed to be the foundational "frame" for constructing Artificial General Intelligence (AGI) systems. By leveraging advanced modules such as the Sleep, Introspection, and Emotional modules, the AGI Framework not only enhances AI capabilities but also ensures a balanced integration of emotional and logical reasoning.

Overview​

The AGI Framework is crafted to bridge the gap between specialized and general-purpose AI, offering a flexible infrastructure for coordinating multiple AI capabilities. It integrates various AI models and tools into a cohesive system, facilitating the development of AGI systems capable of conscious, intent-driven actions and autonomous execution. The framework's innovative architecture supports multi-modal processing, real-time adaptation, and AGI-to-AGI communication, and support for multiple consciousnesses, making it a strategic asset for users seeking to revolutionize their operations. With features like the Model Manager, Consciousness Manager, and Module Router, the AGI Framework ensures seamless data flow and processing, while its privacy-focused design supports secure, on-premises deployment.

Rapid Release Announcement​

Due to significant developments in the AI industry, including the recent release of high-efficiency open-source reasoning models, we have decided to release The AGI Framework prototype earlier than planned. This decision reflects the dynamic nature of the field and the need to establish the framework's vision and goals promptly. While the framework is not yet fully functional, the release includes comprehensive documentation of its architecture and design. We invite contributors to collaborate on completing the implementation in the coming weeks.

AGI Framework Foundation​

We are exploring the establishment of the AGI Framework Foundation, a non-profit organization aimed at promoting the ethical development and deployment of Artificial General Intelligence. Our vision is to harness the power of AGI to create a better future for everyone. Once established, the foundation will aim to work closely with policymakers and economists to draft proposals for legislation to ensure the continued success of the global economy in the face of rapid technological change and advancement. We are currently seeking public interest and support to make this vision a reality.

Key Features​

  • Model Agnostic: Works with any AI model or tool
  • Multi-Modal Processing: Integrates text, images, audio, and sensor data
  • Flexible Integration: Connects to any API or software system
  • Intent-Driven Actions: Facilitates actions guided by inferred or explicitly stated intentions
  • Autonomous Execution: Enables systems to execute strategies independently
  • API Integration: Supports integration with any API or software system
  • Conscious Processing: Provides a conscious intent-driven processing and execution architecture
  • Multi-Consciousness Support: Supports the simultaneous operation of multiple consciousnesses within the same system
  • Open Source: Community-driven development and customization
  • Enterprise Ready: Scales from small applications to large systems
  • Privacy Focused: Supports on-premises deployment and secure environments
  • Advanced Modules: Includes Sleep, Introspection, Emotional, Decision and Rationalization, and Model Manager modules for enhanced functionality
  • Model Manager: Manages connections between LLMs or models and the modules of The AGI Framework.
  • Consciousness Manager: Manages the creation, deletion, and modification of consciousnesses.
  • Model-Tied Modules: Modules can be individually assigned to any given model, allowing for added flexibility and customization for specific use cases.
  • Custom Prompts: Each module can have a custom system prompt that guides its behavior.

How It Works​

Think of the AGI Framework as the frame of a car: it doesn't reinvent the engine (AI models) or wheels (tools), but provides the structure to bring them together. Built on Django and Django REST Framework, the system enables:

  • Seamless communication between different AI models using a standardized English-first protocol
  • Coordinated execution of complex tasks through JSON-structured messaging
  • Real-time adaptation to changing conditions via self-hosted web servers
  • Safe and ethical decision-making with comprehensive monitoring
  • Integration of diverse data sources for environmental awareness
  • AGI-to-AGI communication for collaborative tasks and resource sharing

The framework's modular architecture includes components for sensory processing, intent recognition, planning, emotional processing, and decision-making, all working together to create a cohesive system that can understand, plan, and execute complex tasks while maintaining strict safety standards.

New Modules​

Sleep Module​

  • Acts as a garbage collector for the system state, ensuring optimal performance by clearing unnecessary or outdated information.

Introspection Module​

  • Enhances the AGI's self-awareness by analyzing the system state and proposing or executing actions based on internal evaluations.

Emotional Module​

  • Interprets and updates the AGI's emotional state, facilitating more natural and effective interactions with humans.

Decision and Rationalization Module​

  • Ensures actions are guided by a balanced integration of emotional and logical reasoning, enhancing decision-making capabilities.

Model Manager​

  • Manages connections between LLMs or models and the modules of The AGI Framework.

Module Router​

  • Handles routing of communications between modules, ensuring seamless data flow and processing within the framework. It acts as a central hub for module interactions, determining the appropriate destination for outgoing data.

Focus Module​

  • Manages the prioritization of tasks and goals, ensuring that the most important tasks are completed first.

Summarization Module​

  • To optimize performance, a Summarization Module will condense the conversation log while preserving essential context. This ensures efficient processing and faster response times without sacrificing task understanding.

User Output Module​

  • Handles all types of user outputs (e.g., text, voice, images, videos).
  • Handles all intent permission requests and responses.
  • Normalizes and preprocesses user outputs for downstream modules.

Consciousness Manager Module​

  • Manages the creation, deletion, and modification of consciousnesses.

New Architectural Approach: Running Consciousnesses​

Defining Consciousness and Intelligence​

The AGI Framework adopts a groundbreaking perspective on consciousness and intelligence, applicable to both artificial and biological systems.

Consciousness​

  • Definition: The ability for individual systems to work together in modules and subroutines to perceive the universe and act in a way that helps achieve a goal.

Intelligence​

  • Definition: A measure of how effectively and efficiently the conscious system achieves its goal.

Significance​

These definitions have profound implications for both AI research and human psychology. By providing a unified understanding of consciousness and intelligence, the AGI Framework bridges the gap between artificial and biological systems, offering insights into the nature of cognition and goal-directed behavior.

  • AI Research: This perspective enables the development of more sophisticated AI systems that mimic the collaborative and goal-oriented nature of biological consciousness.

  • Human Psychology: Understanding consciousness and intelligence in this way offers new avenues for exploring human cognition, behavior, and the underlying mechanisms of goal achievement.

The AGI Framework leverages these definitions to create a cohesive system that integrates diverse AI models and tools, facilitating intent-driven actions and autonomous execution.

Running Consciousnesses​

The AGI Framework is transitioning to a new architecture where each module passes around a running conversation log called a consciousness. This consciousness includes the entire history of inputs and outputs, as well as metadata, allowing modules to build on each other's outputs and maintain a comprehensive context.

Benefits​

  1. Context Preservation:

    • Maintains a complete history of interactions, providing each module with full context for decision-making.

  2. Enhanced Learning and Adaptation:

    • Enables the system to learn from past interactions, improving decision-making and adaptability.

  3. Traceability and Debugging:

    • Offers a clear trace of decision-making processes, aiding in debugging and transparency.

  4. Collaborative Processing:

    • Allows modules to build on each other's outputs, leading to more sophisticated results.

  5. Rich Data for Analysis:

    • Provides a wealth of data for identifying patterns and areas for improvement.

Summarization Module​

To optimize performance, a Summarization Module will condense the consciousness context while preserving essential context. This ensures efficient processing and faster response times without sacrificing task understanding.

Multi-Consciousness Support​

The AGI Framework now supports the simultaneous operation of multiple consciousnesses within the same system. Each consciousness functions as a separate running consciousness, sharing a common short-term and long-term memory and a unified knowledge base.

Key Features​

  • Distinct Identities: Each consciousness has its own identity, allowing for differentiation and specialized interactions.
  • Shared Memory: Consciousnesses share short-term and long-term memory, ensuring consistent knowledge and context.
  • Inter-Consciousness Communication: Consciousnesses can communicate with each other through the Output Module, facilitating collaboration and information exchange.
  • Resource Allocation: The Focus Module manages resource allocation and execution priority among consciousnesses, optimizing system performance.
  • Scalability: The number of consciousnesses is limited only by available computational resources, allowing for flexible scaling based on system capabilities.

User Interaction​

Users can interact with consciousnesses globally, individually, or in groups, providing versatile engagement options tailored to specific needs and preferences.

Practical Examples​

  • Collaborative Problem Solving: Multiple consciousnesses can work together to solve complex problems, leveraging shared knowledge and diverse perspectives.
  • Dynamic Adaptation: The system can dynamically adjust to changing conditions and user needs, providing tailored responses and solutions.
  • Resource Efficiency: The Focus Module ensures optimal resource utilization, balancing workload and execution priority among consciousnesses.

Differentiating The AGI Framework from AI Agents like Operator​

The AGI Framework is designed to redefine how artificial general intelligence operates and integrates across applications, APIs, and real-world systems. Unlike tools such as OpenAI's Operator, which operate primarily within the constraints of a browser and web-based interactions, The AGI Framework offers unparalleled flexibility, extensibility, and functionality.

Advantages Over Similar Tools​

Unlike browser-based AI agents (e.g., OpenAI's Operator), the AGI Framework:

  • Runs anywhere, not just in web browsers
  • Integrates with any software or hardware system
  • Supports complex, multi-step tasks
  • Allows full customization of workflows
  • Enables private, on-premises deployment
  • Provides advanced custom logic and reasoning capabilities

Use Cases​

  • Personal Assistant: Helps you with your daily tasks, such as driving, scheduling, reminders, and task completion.
  • Self-Driving Car: Your car drives itself, such as navigating traffic, parking, and avoiding accidents.
  • Industrial Automation: Factories and other industrial environments automate their operations, such as ordering parts, assisting with production, managing inventory, and quality control.
  • Robotics: Robots and other automated systems perform tasks, such as driving, picking up objects, and assembling products.
  • Smart Home: Your home automates itself, such as controlling lighting, temperature, and security according to your preferences.
  • Smart City: Cities automate their operations, such as traffic control, waste management, and emergency response.
  • Smart Grid: Power grids easily simplify load balancing, fault detection, and maintenance in real-time.
  • Smart Anything: Enables any internet-connected device to leverage AGI capabilities, providing human-like task performance across a wide range of applications with potentially superior results.

Future Directions​

  • Development of Multi-Modal AI Architectures
  • Enhanced Reinforcement Learning Models
  • Explainable AI (XAI) Integration
  • Human-AI Collaboration Tools
  • Introspection and Self-Awareness: Further development of the introspection module to enhance AGI's self-awareness and autonomous decision-making capabilities.

On the Future of Jobs​

Rather than eliminating jobs, we firmly believe the AGI Framework will create new opportunities. History shows that technological advances, while disrupting some industries, ultimately generate more jobs and economic growth. Our framework empowers people to tackle previously impossible tasks, freeing them to focus on what humans do best - creative thinking and innovation. While machines excel at repetitive or dangerous work, human creativity remains irreplaceable. We're excited to help unlock human potential in new ways, enabling solutions to problems we couldn't solve before. The AGI Framework isn't about replacing workers - it's about augmenting human capabilities and creating new possibilities.

Key Challenges & Solutions​

  1. Safety: Comprehensive safety protocols and monitoring systems
  2. Scalability: Modular architecture and microservices approach
  3. Privacy: Secure communication protocols and sandboxed environments
  4. Integration: Flexible middleware solutions for legacy system compatibility
  5. Contextual Awareness: Advanced cross-modal attention mechanisms and robust data preprocessing
  6. Alignment and Safety: Comprehensive safety architecture and operational oversight tools

Safety​

Like any powerful technology, the AGI Framework must be used responsibly. While we have implemented comprehensive safety features and security measures, we recognize there may be concerns about potential misuse.

The framework itself is not inherently dangerous - it is a tool, similar to other powerful technologies like the internet or nuclear power. The key is responsible usage and oversight. While we cannot guarantee the framework will never be misused, we have built in multiple safety mechanisms to help prevent harmful applications.

The AGI Framework prioritizes safety and ethical considerations through:

  • Built-in principles of transparency, accountability, and fairness
  • Support for diverse applications from personal to industrial use
  • Clear responsibility boundaries and guidelines

While the framework provides robust safety features, users should note:

  • It is a tool that must be used responsibly
  • The framework acts as a neutral facilitator of AI capabilities
  • It does not make ethical or political judgments

Ultimately, responsibility for ethical usage lies with:

  • End users implementing the framework
  • Developers maintaining the underlying models and tools
  • Organizations deploying AGI systems

We take safety extremely seriously and have detailed our comprehensive safety architecture and guidelines in safety. The AGI Framework aims to advance beneficial AI development while maintaining robust safeguards against misuse.

Scalability​

The AGI Framework is designed to be scalable and modular. It is designed to be used in a wide range of applications, from personal assistance to industrial automation. It achieves this by using a microservices architecture and a modular approach to development. The AGI Framework does not have a single point of failure and is designed to be able to scale to any size of application. Please see scalability for more details.

Privacy​

The AGI Framework is designed to be privacy-focused and secure. It achieves this by using a secure communication protocol and a modular approach to development. This modular approach provides a high level of security and privacy by design, but ultimately the responsibility for ensuring that the AGI Framework is used in a privacy-focused way rests with the end-user. You should always use the AGI Framework in a way that is consistent with your privacy policies and practices. Please see privacy for more details.

Integration​

The AGI Framework is designed to be integrated with any software or hardware system. This is accomplished by allowing virtually any form of input and output to be used with the AGI Framework, including text, images, audio, sensor data, raw binary data, API requests and responses, and any other form of data. In order to achieve this tight integration, the AGI Framework is designed to be a middleware layer between the end-user and the AI models and tools. Please see integration for more details.

Contextual Awareness​

The AGI Framework is designed to be aware of the context of the user and the environment. This is accomplished by using a modular approach to development that allows for the integration of diverse data sources and the use of advanced cross-modal attention mechanisms. Please see modules and the complete AGI Framework documentation for more details.

Alignment and Safety​

The AGI Framework is designed to be aligned with the values and goals of the user, as well as utilize the included safety features of the underlying models and tools. The framework does not make any ethical or political judgments. However, the inclusion of the emotional module as well as other advanced processing modules are designed to help overcome some of the limitations of models that are less ethically aligned. Please see safety for more details.

Installation and Setup​

The AGI Framework will soon be available as a unified installer that provides a seamless setup experience across all platforms:

  • Unified Installer: Bundles all dependencies including:

    • Python and pip
    • PostgreSQL
    • Ollama
    • Django and Django REST Framework
    • Node.js and npm
    • All required Python and JavaScript packages

  • Electron App: A modern desktop application providing:

    • Intuitive graphical interface
    • System tray integration
    • Background services management
    • Automatic updates

Web Interface​

Our new Next.js frontend, built with shadcn/ui, provides a modern and intuitive user experience:

  • Custom Module Management:

    • Add new modules through a simple form interface
    • Configure module settings and parameters
    • Monitor module status and performance
    • Visual module dependency mapping

  • Dashboard Features:

    • Real-time system monitoring
    • Module performance metrics
    • Consciousness state visualization
    • Task queue management

  • Responsive Design:

    • Works seamlessly across desktop and mobile
    • Dark and light theme support
    • Accessible interface
    • Intuitive navigation

Module Development​

The new web UI simplifies module creation and management:

  • Module Creation Form:

    • Define module parameters
    • Set input/output specifications
    • Configure model bindings
    • Specify system prompts

  • Module Testing Interface:

    • Real-time module testing
    • Performance monitoring
    • Debug logging
    • Error tracking

Getting Started​

  1. Read our Code of Conduct

  2. Fork the Repository: Start by forking the main repository to your Codeberg account.

  3. Clone the Repository: Clone your fork to your local machine using:

    git clone https://codeberg.org/yourusername/AGI-Framework.git

  4. Set Up the Environment:

    • Install the required system dependencies:

      sudo apt install python3-pip python3-venv

    • Navigate to the project directory:

      cd AGI-Framework

    • Create a virtual environment:

      python3 -m venv venv

    • Activate the virtual environment:

      • On Windows: 
        venv\Scripts\activate
      • On macOS/Linux: 
        source venv/bin/activate

    • Install the required dependencies:

      pip install -r requirements.txt

    • Set up environment variables by creating a .env file and adding necessary configurations (refer to example.env for guidance).

    • By default, the framework uses a PostgreSQL database. If you want to use a different database, you will need to modify the settings.py file.

    • Configure the database and then run the following commands:

      python manage.py makemigrations
      python manage.py migrate
      python manage.py runserver

      If you are having trouble with the migrations, you can try the following commands to set up PostgreSQL from scratch:

      # Update package lists
      sudo apt update

      # Install PostgreSQL
      sudo apt install postgresql

      # Switch to the postgres user
      sudo -u postgres psql

      # Create a new database
      CREATE DATABASE agi_framework;

      # Create a new user with a password
      CREATE USER myuser WITH PASSWORD 'changeme';

      # Set the client encoding to UTF-8
      ALTER ROLE myuser SET client_encoding TO 'utf8';

      # Set the default transaction isolation mode to read committed
      ALTER ROLE myuser SET default_transaction_isolation TO 'read committed';

      # Set the timezone to UTC
      ALTER ROLE myuser SET timezone TO 'UTC';

      # Grant all privileges on the database to the new user
      GRANT ALL PRIVILEGES ON DATABASE agi_framework TO myuser;

      # Grant usage and create on the public schema
      GRANT USAGE, CREATE ON SCHEMA public TO myuser;

      # Grant all privileges on all tables in the public schema to the new user
      ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON TABLES TO myuser;

      # Set the owner of the database to the new user
      ALTER DATABASE agi_framework OWNER TO myuser;

      # Exit the PostgreSQL prompt
      \q

  5. Read our Contribution Guidelines

  6. Create a new branch for your contribution.

  7. Submit a pull request.

Licensing​

The AGI Framework is released under the MIT License, allowing:

  • Unrestricted use in both personal and commercial applications
  • Complete freedom to modify and distribute the software
  • Maximum innovation potential through minimal restrictions
  • Equal access to AGI technology for all stakeholders

See the LICENSE file for details.

Packaging and Distribution​

The AGI Framework is packaged as an Electron application, providing a unified installer that bundles all necessary dependencies, including Python, Pip, Django, PostgreSQL, and Ollama. This ensures a smooth user experience with a modern UI and minimal configuration required. The community licensed version of our prototype runs on both an Electron app and a web interface, offering flexibility and ease of use. Although the commercial version shares the same features as the community version, it supports multiple users, catering to larger organizations and enterprises.

Contributing​

This is an open-source project. Contributions are welcome through:

  • Code contributions
  • Feature suggestions
  • Bug reports
  • Documentation improvements

We are looking for contributors to help us build this project. If you are interested, please see contributing for more details.

  • Bitcoin: bc1q896spt439ghna42f5mq5eynadkt6w9936dkc9f
  • Litecoin: ltc1qqxx67dprszdxa7xfdumhydagh45cmp56yx5sed
  • Ethereum: 0x99F58f8430dc495DA404d88FE47A0913855d30bB

If you would like to support the development of The AGI Framework, please see donate for more details.

All donations are greatly appreciated and will be used to support the development of The AGI Framework and the AGI Framework Foundation.

Updates​

This project's documentation and introduction will be updated frequently as we continue to build out the framework. Please check back regularly for the latest updates. Major changes between versions will be noted in the changelog file.

Contact​

For more information, please contact us at [email protected].

Acknowledgements​

This project is built on the shoulders of giants. We would like to thank the following projects for their inspiration and contributions:

License​

This project is licensed under the MIT License. See the LICENSE file for details.

Warranty​

This project is provided "as is", without any warranty. Use at your own risk.

Copyright [2025] [The AGI Framework]