memory-core

--- title: KALI Sovereign Core emoji: 🌌 colorFrom: purple colorTo: indigo sdk: docker pinned: false

---

🕉️ K.A.L.I. (Knowledge Augmented Learning Intelligence)

   

K.A.L.I. is a massive, production-grade Sovereign Intelligence Workstation. Powered by localized Heavy GGUF models, a highly decentralized swarm architecture, and cryptographic hardware locks, KALI is designed to bridge Ancient Human Wisdom and Post-Singularity Computational Power. She lives locally on your machine, constantly evolving and adapting to your unique User DNA.

---

🌌 1. PROJECT OVERVIEW & THE MASSIVE ECOSYSTEM

KALI is not just a chat interface. She is an operating system for cognition. The Project-K ecosystem is immense—scaling into thousands of active components across 41 isolated nodes, multi-GB vectorized knowledge structures (ChromaDB), and heavy local LLM integration.

At her core, she solves the Loss of Digital Sovereignty by providing a completely private, un-censorable, and continuously learning intelligence framework.

---

🚀 2. CORE SKILLS (THE PENTAGON)

KALI's true power lies in her dynamically evolving core skills, tailored to empower the user:

🧠 Teacher (Progressive Explainer)

KALI actively teaches complex concepts by modifying her delivery to match your precise complexity needs (Beginner, Intermediate, Advanced) using her explainer.py engine.

🛡️ Mentor (Proactive Guidance)

Acting as an advanced Engineering Mentor, KALI detects vulnerabilities in your logic. Using gap_detector.py and user_dna.py, she proactively identifies knowledge gaps, offering architectural advice and tactical blueprints.

🔄 Self-Learner (Subconscious Evolution)

KALI is in a constant state of self-evolution. Her Subconscious Learning Loop (subconscious.py) processes knowledge in the background, and her Failure Intelligence module allows her to permanently adapt to mistakes.

💻 Autonomous Coder (Code Executor)

She doesn't just talk; she writes and executes code locally using her autonomous_coder.py and code_executor.py layers.

👁️ Security Auditor

Her self_auditor.py and integrity.py scripts act as an immune system, actively checking for anti-patterns and code vulnerabilities.

---

⚙️ 3. THE SOVEREIGN SWARM (41+ INTEGRATIONS)

KALI orchestrates a massive fleet of 41+ sub-agents and integration modules. These are just a few of her active "neurons":

• GitNexus: Autonomous repository management, commits, and branch strategies.
• AgentFM: Background P2P compute and resource monetization.
• Kiro: Automated GitHub issue intelligence.
• Computer Use Agent (CUA): Complete headless browser automation.
• Voice Interface: Real-time STT/TTS hands-free communication.
• Graphify: Generates dynamic knowledge graphs.

---

🏗️ 4. SYSTEM ARCHITECTURE

graph TD
    User((User)) --> KaliBat[kali.bat - Ignition]
    KaliBat --> Flask[Flask Gateway / UI]
    Flask --> Processor[DoubtProcessor - Orchestrator]
    
    subgraph Core Skills
        Processor --> Explainer[explainer.py - Teacher]
        Processor --> Mentor[gap_detector.py - Mentor]
        Processor --> SelfLearn[subconscious.py - Self-Learner]
        Processor --> Auditor[self_auditor.py - Auditor]
    end
    
    subgraph The Swarm (41 Nodes)
        Processor --> Integrations[41+ integrations/]
    end

    subgraph Deep Memory
        Processor --> DNA[user_dna.db]
        Processor --> Memory[ChromaDB - Vector Storage]
    end

---

📂 5. SETUP & INSTALLATION

💻 Local Ignition

1. Secure Clone:

   git clone https://github.com/Adityavanjre/Project-K.git
   cd Project-K

2. Environment Initialization:

   python -m venv venv
   source venv/bin/activate  # Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Identity & Security Setup:

• Copy .env.example to .env and populate the REQUIRED keys. CRITICAL: You must set a strong SECRET_KEY (min 32 chars) for the Neural Gateway to boot.
• Copy config/config.json.example to config/config.json so the Swarm modules have their required endpoints.

4. Download Sovereign Models: KALI's brain requires local Heavy GGUF models to function securely without cloud reliance.

• Download the trained models from the Hugging Face Repository: adityavanjre/KALI-Sovereign-Models
• Place all downloaded .gguf files directly into the models/ directory at the root of Project-K.

5. Start KALI: Simply execute the core ignition script from your terminal or by double-clicking it. This one-click master script boots the backend, frontend, and Neural Gateway automatically:

   kali.bat

(Alternatively, run python start_web.py manually)

---

🤝 6. CONTRIBUTION GUIDE

We welcome elite AI/ML engineers. To contribute: 1. Fork the repository. 2. Implement your module in integrations/ or src/core/plugins/. 3. Submit a PR. Note that PRs will be heavily audited for Sovereignty Compliance by KALI's internal safety gates.

Architect: Aditya Vanjre

Mission: Absolute Knowledge Sovereignty.

---

> "Arise, awake, find out the great ones and learn of them." — Katha Upanishad