CLO — Cross-LLM Context Orchestrator

Deterministic Reasoning Portability for the Era of Heterogeneous Intelligence.

CLO (Cross-LLM Orchestrator) is a specialized browser extension designed to eliminate context loss when switching between different LLM platforms. It acts as a "Reasoning State Virtual Machine", capturing the cognitive state of a conversation and enabling it to be "rehydrated" into any other model seamlessly.

---

✨ Features

• 🔄 Reasoning Portability: Transfer complex tasks between ChatGPT, Claude, Gemini, and Grok without repeating yourself.
• 🧠 Context Extraction: Automatically parses constraints, decisions, and artifacts from your conversation using a canonical JSON schema.
• ⚡ Zero-Friction Rehydration: Single-click "Inject" automatically compiles and adapts your state into the next model's input field.
• 📉 Semantic Compression: Intelligently compresses conversation history to save tokens while preserving critical reasoning paths.
• 🔗 Provenance Tracking: A "git log" for your thoughts — know exactly which model made which decision.
• 💎 Premium HUD: Minimalist, glassmorphism-inspired UI that stays out of your way until you need it.

---

🛠 Quick Start

1. Installation

Currently in Developer Preview. To install manually:

1. Clone the Repo:

   git clone https://github.com/your-repo/clo-extension.git
   cd clo-extension

2. Install Dependencies:

   npm install

3. Build the Project:

   npm run build

2. Loading into Chrome/Brave/Edge

1. Open your browser and navigate to chrome://extensions/.
2. Enable Developer mode (toggle in the top-right corner).
3. Click Load unpacked.
4. Select the dist folder (created inside your project directory after running the build).

---

📖 Usage Guide

1. Capture: Open any supported LLM (ChatGPT, Claude, or Gemini). Start your conversation as usual.
2. Monitor: Notice the CLO HUD in the bottom-right corner. It tracks turns, detected models, and reasoning primitives in real-time.
3. Export/Switch: When you reach a point where you want to switch models (e.g., from GPT-4o to Claude 3.5 Sonnet):
   • Simply open the new platform.
   • Click ⚡ Inject in the HUD.
   • CLO will automatically compile the reasoning state and paste it into the prompt box.
4. Continue: Hit Enter, and the new model will pick up exactly where the last one left off, aware of all previous constraints and decisions.

---

🏗 Architecture

graph TD
    A[LLM Platform DOM] -->|Capture| B(Platform Interceptors)
    B -->|Structured Data| C{Background Orchestrator}
    C -->|Persistence| D[(IndexedDB State Store)]
    C -->|Stats/UI| E[Premium HUD / Popup]
    C -->|Logic| F(Serializer + Compressor)
    F -->|Compiled Prompt| G[Injection Engine]
    G -->|DOM Write| H[Target LLM Input]

Loading

Core Primitives

Component	Responsibility
Interceptors	Platform-specific DOM observers (MutationObserver)
Serializer	Mapping raw text to the ReasoningState schema
Compressor	Semantic folding of long histories using model-aware logic
Injector	Adapting state into model-specific prompt frames

---

📋 Platform Support Matrix

Platform	Status	Model Detection	Capture	Injection
ChatGPT	✅ Active	✅	✅	✅
Claude	✅ Active	✅	✅	✅
Gemini	✅ Active	✅	✅	✅
Grok	🟡 Beta	Minimal	🚧	✅

---

🔬 Contributing

We're building the future of deterministic reasoning. Contributions are welcome!

1. Fork the repository.
2. Create your feature branch (git checkout -b feature/amazing-feature).
3. Commit your changes (git commit -m 'Add amazing feature').
4. Push to the branch (push origin feature/amazing-feature).
5. Open a Pull Request.

---

📄 License

Distributed under the MIT License. See LICENSE for more information.

---

Built with ❤️ for the LLM community.

By - Anurag