Morphic Lattice AI

Morphic Lattice AI is a browser-native AI computation lab for discovering unusual engineering interventions. Instead of behaving like a normal chatbot, it turns a problem brief into a living causal lattice and runs a local evolutionary search over possible counterfactual actions.

The result is an interactive instrument that helps engineers explore hidden pressure points, unstable dependencies, novelty pockets, and risk-aware system maneuvers.

What Makes It Unique

Most AI products summarize text, answer questions, classify data, or generate content. Morphic Lattice AI is different: it simulates a problem as a dynamic system and searches for interventions inside that system.

The product combines:

• Dynamic causal graph simulation
• Local evolutionary search
• Multi-objective scoring
• Real-time canvas visualization
• Web Worker based computation
• Exportable AI-generated intervention hypotheses

It does not require an API key, cloud model, database, or external dependency. The intelligence comes from the local simulation and search engine.

Core Features

• Live causal lattice with nodes, links, pressure, signal, entropy, trust, and novelty
• Search modes for balanced, high-novelty, or conservative exploration
• Adjustable compute budget, population size, risk aversion, and novelty bias
• Ranked intervention ideas with utility, novelty, stability, and risk scores
• Web Worker execution so the UI remains responsive during computation
• Export button for saving experiment results as JSON
• Built-in local static server
• Node test suite for the simulation kernel

Demo Scenario

The default scenario asks the engine to explore:

Design a self-repairing autonomous data center that reduces thermal failures, power spikes, and operator blind spots during extreme load.

You can replace the brief with any complex engineering problem, reset the lattice, and run a fresh search.

Tech Stack

• HTML
• CSS
• JavaScript ES modules
• Canvas API
• Web Workers
• Node.js local server
• Node test runner

Project Structure

morphic-lattice-ai/
  index.html
  package.json
  server.js
  src/
    kernel.js
    main.js
    styles.css
    worker.js
  tests/
    kernel.test.js

Getting Started

Clone or download the repository, then run:

npm run dev

Open the app in your browser:

http://localhost:5173

On Windows PowerShell, if npm is blocked by execution policy, use:

npm.cmd run dev

Run Tests

npm test

On Windows PowerShell:

npm.cmd test

How It Works

1. The app creates a seeded causal lattice with different node roles such as sensors, actuators, constraints, memory, operators, and shadow states.
2. A Web Worker evolves candidate interventions in the background.
3. Each candidate is scored across utility, novelty, stability, and risk.
4. The best candidates influence the lattice state over time.
5. The UI renders the changing graph and displays ranked intervention hypotheses.

Example Intervention Output

The engine produces ideas such as:

• Phase Braid on Sensor
• Constraint Bloom on Operator
• Entropy Clamp on Memory
• Thermal Mirror on Actuator

Each intervention includes a short action description and normalized scoring rationale.

Why This Matters

Complex engineering systems often fail because teams optimize the visible parts of the system while missing hidden dependencies. Morphic Lattice AI gives engineers a way to explore strange but plausible system changes before committing to real-world experiments.

It is designed for:

• Systems engineering
• Infrastructure design
• Reliability research
• Autonomous operations planning
• AI-assisted product ideation
• Experimental decision support

GitHub Upload

If this folder is not already a Git repository:

git init
git add .
git commit -m "Build Morphic Lattice AI prototype"
git branch -M main

Create a new empty repository on GitHub, then connect and push:

git remote add origin https://github.com/YOUR_USERNAME/morphic-lattice-ai.git
git push -u origin main

Roadmap

• Add WebGPU acceleration for larger graph relaxation
• Add a WASM simulation kernel for faster rollouts
• Add saved experiments and replayable traces
• Add CSV/JSON imports for real system maps
• Add scenario templates for data centers, robotics, supply chains, and cyber defense
• Add optional LLM explanations for top-ranked interventions

License

MIT License. See LICENSE.