LYRN-AI Cognitive Architecture - v4.2.10

LYRN (Live-reasoning & Structured Memory) is a highly modular, professional-grade GUI for interacting with local Language Models. It is designed from the ground up for efficiency, accessibility, and genuine cognitive continuity. It features advanced job automation, live system monitoring, a dynamic prompt building system, and a robust, file-based architecture for memory and inter-process communication.

Core Philosophy

LYRN is built on a philosophy that diverges significantly from mainstream LLM development. Instead of relying on ever-larger models and prompt injection, LYRN emphasizes structured, live memory to achieve genuine continuity, identity, and context. The core tenets are:

• Efficiency and Accessibility: The primary goal is to create a powerful AI cognition framework that is lightweight enough to run on standard consumer hardware.
• Structured Memory over Prompt Injection: All core context—personality, memory, goals—lives in structured text files and memory tables. The LLM reasons from this stable foundation rather than having it repeatedly injected into a limited context window.
• Simplicity and Robustness: The architecture is inspired by the simplicity of 1990s text-based game parsers. The framework's job is to be a robust, simple system for moving data; the LLM's job is to do the heavy lifting of reasoning.

For a deeper dive into the project's vision and architectural principles, see AGENTS.md.

Major Features

✅ Modern, Responsive UI: Built with CustomTkinter for a professional look and feel. Asynchronous initialization ensures the UI loads instantly and never freezes during model operations.

✅ Live System Monitoring: Real-time gauges for CPU, RAM, Disk, and VRAM usage.

✅ Advanced Theming Engine: Includes a live, in-app Theme Builder to create, modify, and save custom color themes.

✅ Dynamic System Prompt Builder: A powerful two-panel UI gives you granular control over every component of the system prompt. - Dynamically create, edit, and delete prompt components. - Toggle components on or off. - Reorder components with arrow keys. - Customize block wrappers and content for ultimate flexibility.

✅ Job & Tool Automation: A powerful automation system with: - A Job Manager to create and run complex, multi-step jobs. - A Scheduler with a full calendar view to run jobs at specific dates and times. - A Cycle Manager to create and execute custom, multi-step cognitive cycles. - Support for the gpt-oss "harmony" format for defining tools, allowing the AI to be extended with new capabilities.

✅ Structured, Long-Term Memory: - Episodic Memory: Saves each chat interaction as a structured, searchable file. - Topic Indexing: A background process that identifies and indexes topics from conversations to build a personalized web of meaning over time. - Quote System: Right-click any text in the chat to "Quote to Context" for easy reference.

✅ Chat Persistence & Control: - Chat history is automatically saved and reloaded between sessions. - Fine-grained control over how much chat history and which memory "deltas" are injected into the context for each turn.

✅ Full Model Control: - Save and load model configuration presets for one-click setup. - In-app controls for all key generation parameters (temperature, top_p, top_k, etc.). - Stop generation at any time with a dedicated button.

Screenshots

Dashboard

System Prompt Builder

Job Manager

Chat Settings

Model Settings

Performance Metrics

Quick Start

1. Install Dependencies:

   pip install -r dependencies/requirements.txt

2. Place Models:

   • Create a folder named models in the root directory.
   • Place your GGUF-format model files inside the models/ folder.

3. Run Application:

   python lyrn_sad_v4.2.10.pyw

4. First Launch Setup:

   • On the first launch, the "Model Settings" window will appear.
   • Select your model from the dropdown, adjust parameters if needed, and click "Load Model".
   • The application auto-generates settings.json to store all configurations.

Key Files

Core Application

• lyrn_sad_v4.2.10.pyw - The main GUI application file.
• settings.json - Auto-generated configuration file for model settings, paths, and UI preferences.
• automation/ - Contains all background watcher scripts and configurations for autonomous operation.
• build_prompt/ - Contains all modular components for building the system prompt.
• themes/ - Directory containing JSON theme files.
• dependencies/requirements.txt - A list of all required Python packages.

Documentation

• README.md - This guide.
• build_notes.md - A detailed, version-by-version change log of all updates.
• AGENTS.md - Instructions and architectural rules for AI agent development on this codebase.

Troubleshooting

Common Issues

• Application doesn't start: Ensure all dependencies from dependencies/requirements.txt are installed.
• Model won't load: Check the console output. Ensure the model path is correct in the "Model Settings" popup and the file is not corrupt.
• Missing folders: Most required folders (like models/) are created automatically on startup. If you encounter issues, deleting settings.json will trigger the first-time setup again.
• Automation not working: Ensure the relevant watcher scripts are running. You can check your system's task manager for multiple python processes.

License

This project is licensed under a custom source-available license. See the LICENSE file for details. The key points are:

• The software is provided "as-is".
• You are free to use, modify, and redistribute it for non-commercial purposes.
• Commercial use requires express written permission from the copyright holder (LYRN-AI).