Project T180: The Meta-AI Workspace

Hyper View Manifest: This workspace represents the convergence of desktop interaction, structured memory, and autonomous agency. It is the proving ground for the Klipper SDK and its evolution into a Meta-AI Orchestrator.

🌌 Vision

Project T180 is not just a collection of tools; it is a Cognitive Operating Environment. By bridging the gap between ephemeral user actions (clipboard, shell) and structured long-term memory (SQL), it enables AI agents to observe, learn, and eventually predict user intent.

The core philosophy is built on Five Generations of evolution:

1. Interop: Connecting to the OS (Clipboard/D-Bus).
2. Observation: Structuring raw data into patterns.
3. Capability: Defining tools for agents to act.
4. Agency: creating specialized personas (Planner, Coder, Analyst).
5. Orchestration: Meta-cognitive management of multi-agent workflows.

---

🏛 Architecture Map

The workspace is divided into four distinct planes of existence:

Plane	Component	Description	Technology
The Brain	klipper_sdk	The core intelligence library. Handles ETL, Agent logic, and Orchestration.	Python 3.8+
The API	backend	The nervous system. Exposes "The Brain" to the outside world via REST endpoints.	FastAPI
The Face	frontend	The user interface. Visualizes thought processes, plans, and memory.	React + Vite
The Memory	db	The hippocampus. Stores structured history and learned patterns.	Dolt / MySQL

graph TD
    User((User)) -->|Interacts| UI[Frontend]
    User -->|Copy/Paste| Klipper[KDE Klipper]
    
    UI -->|HTTP| API[Backend API]
    
    subgraph "The Brain (Klipper SDK)"
        API --> Orchestrator
        Orchestrator --> Agents
        Agents --> Tools
        Klipper <-->|D-Bus| Client[SDK Client]
    end
    
    Client -->|ETL| DB[(Dolt DB)]
    Agents -->|Query| DB

Loading

---

🧭 The Five Generations

This project is currently transitioning from Generation 2 (Observation) to Generation 5 (Orchestration).

• Gen 1 (Foundation): Raw access to Klipper history. [✅ DONE]
• Gen 2 (Observer): Star-rating logic, frequency analysis, and basic ETL. [✅ DONE]
• Gen 3 (Tools): Standardized tool interfaces for code execution and file IO. [✅ DONE]
• Gen 4 (Agents): Role-based execution (e.g., "The Analyst" vs "The Coder"). [✅ DONE]
• Gen 5 (Orchestrator): KickLang blueprint execution and dynamic workflow adaptation. [✅ DONE]

---

🚀 Global Quick Start

To bring the entire system online:

1. Initialize Memory

# (Optional) Navigate to db to handle schema
cd db
dolt init
dolt sql < doltdump.sql
dolt sql < seed_minimal.sql

Ensure your database is running and accessible.

dolt sql-server -P 4448

Server will listen on mysql://localhost:4448

2. Awaken the Brain (Backend)

Start the FastAPI server to expose the SDK capabilities.

cd backend
pip install -r requirements.txt
python main.py

Server will listen on http://localhost:8000

3. Open the Eyes (Frontend)

Launch the dashboard to interact with the system.

cd frontend
npm install
npm run dev

UI will be available at http://localhost:5173

---

"The goal is not just to automate the task, but to understand the intent behind it."