The Moset Ecosystem
From the physical-mathematical bases of our Artificial Intelligence to code injection without the barrier of forced English. This is how you master the Moset tools.
1 Quantum Physics and Sovereign Inference
Unlike traditional LMs that base their temperature on the statistical thermal concept, the Moset Sovereign AI Engine manipulates abstract temperature using the Born Rule.
The Amplitude Metric $|0\rangle$
In the Moset IDE, you will see an Amplitude slider that determines the exact entropy. That probabilistic amplitude $\alpha$ decays into the "Temperature" ($T \approx |\alpha|^2$) that is injected into the native processing node or cloud APIs. This grants us near-perfect determinism for programming.
🧬 Train of Thought: The <think> Block
Regardless of whether you connect to third-party models or a local model on your own machine, Moset forces the model (at the Hidden System Prompt level) to perform its deep computation within a structural XML darkroom. First it reasons, then it exposes. Zero hallucinations.
<think>
1. Analyze topological lexer.
2. Identify syntax failure in Base-1.
3. Plan correction using U-AST.
</think>
Here is your corrected code...
2 The U-AST .et Dialect
The dialect eliminates the use of curly braces and semicolons, using precise topological indentation accompanied by visual Macros. But the revolution is not aesthetic, it is about intrinsic translation.
Human Base-1
Arrays start at [1], aligning semantics with human reading, applying a natural 'Zerocost Abstraction' in the backend towards binary memory.
Autonomous U-AST Translation
A developer in Spain sees the code in Spanish, and if the .et file is opened by a Japanese developer, the AST live-decodes the tokens into Japanese.
3 Moset IDE (Graphic Environment)
The native graphical environment compiles into a secure system, ensuring minimal RAM consumption while you inject the power of the core engine.
1. Universal Translation Lens
Activate the "Lens" in settings to visually project keywords into your language (if, while, struct) without altering the original .et file. It is an augmented reality layer for global programmers.
2. GGUF Metadata Editor
Open any AI model (.gguf) and access the tensor and metadata editor. Modify the model's name, author, or parameters directly from the graphical interface without using external tools.
3. Sovereign Agent: Act Mode
The AI agent now uses a "Group Approval" flow. It proposes a complete multi-file plan and you authorize them in block. Includes Vigilante protection for secure executions.
4. Drag & Drop Context
Inject instant knowledge into the chat by dragging files directly from your desktop. The agent will read the content and add it to its internal reasoning in a single motion.
4 Terminal (Moset CLI)
For managing remote servers without a graphical interface, use our minimalist CLI. A single ultra-fast binary invoked from your preferred command line.
# Transpile to bytecode for Server $ moset build vps_controller.et
# Execute live code hot $ moset run security_tests.et
5 WASM Execution Engine
Executing .et code in web environments does not rely on external servers. We have implemented an asynchronous WebAssembly (WASM) Engine that compiles and executes the dialect directly in the client's browser.
Isolated Asynchronous Execution
The engine evaluates the AST (Abstract Syntax Tree) in separate threads using Web Workers, preventing UI blocking even in complex inference calculations.
Zero-Trust Security
All code runs in a strict sandbox. The WASM engine has no access to the DOM or browser APIs unless explicitly declared through Moset's security bridges.
