Projects
A portfolio of apps, frameworks, engines, demonstrators, and research systems I build under the Raven Forge name — each built around disciplined architecture, modular design, and long-term technical integrity.
Rune Scribe
Turn your words into Norse runes — and back again. A private, on-device substitution-cipher app for macOS and iOS. Type a message, watch each character inscribe itself into a unique rune, and share the result; hand a friend your decoder file and they can read it, while everyone else sees only runes.
- Encode & decode an 89-character set into Unicode runes
- A library of labelled cipher / decipher pairs
- Share a decoder file so a friend can read your message
- Optional 4-digit app lock (PBKDF2, rate-limited)
- On-device only — no servers, accounts, or tracking
- Universal app for macOS 15+ and iOS 17+
Mirmir
A macOS video upscaling app. Choose a source video, pick an upscale strategy, tune quality and output settings, and run a deterministic, Metal-accelerated upscale job — built with SwiftUI, AVFoundation, Metal, and VideoToolbox on the Forsetti Framework.
- Deterministic, Metal-accelerated upscaling
- AVFoundation ingest & export pipeline
- VideoToolbox model-asset readiness
- Configurable quality and output settings
- Built on the Forsetti Framework (macOS 15.4+)
Skald
A macOS app that batch-converts documents into clean Markdown or structured JSON. Point it at a folder and Skald turns PDFs, Office/RTF/HTML, spreadsheets, images (OCR), and structured data into readable, pipeline-ready output — built on the Forsetti Framework.
- Batch folder-to-folder conversion
- Markdown with heading, list & paragraph reconstruction
- Structured JSON model for automation & LLM pipelines
- PDF, Office, RTF, HTML, CSV, JSON, XML, plist & more
- Image OCR via Vision; verbatim code & log blocks
Forsetti Jamf Pro
A macOS and iOS operations app for Jamf Pro, built as a Forsetti consumer application. Search computers and mobile devices, run diagnostics, track deployments, manage prestages and permissions, and support technicians from one native app.
- Computer & mobile-device search
- Deployment tracker & prestage director
- Reports and permissions matrix
- Support-technician & scanner tools
- Jamf API authentication & diagnostics
- Native macOS & iOS (Forsetti Pattern B)
Forsetti Framework
A compatibility-governed, entitlement-aware modular runtime framework for native applications. Forsetti gives host apps a consistent way to discover, validate, unlock, and activate feature modules through explicit manifests and contracts — with capability-scoped services and structured UI contributions — delivered as native platform editions plus a governance edition for AI coding agents.
- Compatibility-governed module model with manifest validation
- Entitlement-aware module locking and activation policy
- Capability-scoped services and framework-mediated messaging
- Native C++20 edition for Windows 11 (MSVC, WinHTTP, DPAPI)
- Apple-native Swift & SwiftUI edition for macOS & iOS
- Agentic Edition: governance for AI coding agents
Yggdrasil World Engine
A world-building engine for creating interconnected, persistent digital environments. Yggdrasil manages world graphs, real-time state synchronization, and scalable environment persistence across platforms.
- Interconnected world graph architecture
- Real-time state synchronization engine
- Persistent world management
- Scalable multi-environment deployment
- Dynamic content loading and streaming
- Event-driven world interaction system
Sigil
My first working Raven Forge demonstrator of the ASH Model applied as deterministic software. A proof of concept — not a consumer release — that turns a canonical identity into reproducible symbolic state, sigil geometry, narrative lore, and portable, self-validating artifacts.
- Deterministic ASH / WRW generation pipeline
- 9-bit GF(2) ASH state with a guarded codeword orbit
- 48-sided 3D sigil mesh and 2D vector exports
- Schema-versioned PersonalCodex identity artifact
- Self-validating exports via the Aeostara health layer
- Forsetti-hosted, capability-governed modules
Aeostara
A self-healing JSON configuration engine. Aeostara observes live configuration against a declared desired state, diagnoses drift, and executes policy-gated, verifiable recovery with rollback and a full audit trail — built on the ASH Pattern System, with native Windows, macOS, and iOS realizations.
- Diagnosis-first drift detection, not blind JSON diffing
- Policy-gated, verifiable recovery with backup & rollback
- Full diagnostic and audit chain
- Native C++20 edition for Windows
- Apple-native macOS & iOS realizations
- Conforms to the ASH Pattern System
Master Control Orchestration Layer Server
A Windows-native LAN MCP gateway host. MCOS exposes one advertised MCP endpoint for trusted-LAN clients, supervises MCP-server and sub-agent worker pools, and distributes onboarding profiles and CLU/Forsetti governance bundles — with browser and WinUI maintainer surfaces.
- Native HTTP.sys MCP gateway for trusted-LAN clients
- DNS-SD / mDNS discovery and per-client onboarding profiles
- Managed MCP-server & sub-agent worker pools
- CLU / Forsetti governance bundle distribution
- Job Object process containment and autoscaling
- MSI packaging — Windows 11 / Server 2022 (x64)
ASH Model
The Adinkra-Stabilized Hypercube Model — the deterministic framework at the root of every Raven Forge project, and of my writing. It is built on a proven 9-bit hypercube mathematics, then extended into specified deterministic computation and interpretive simulation-theory research.
- Proven finite core: a 9-bit hypercube of 512 states
- Canonical [9,4,4] doubly-even code with 16 codewords
- Adinkra Q8/C8 quotient and exact N=8 Garden matrices
- Idempotent code-orbit projection and guaranteed one-bit correction
- Deterministic image/video state mapping and bounded branching
- Honest scope: proven math, specified computation, interpretive research
ASH Pattern System
An applied framework for building self-healing, self-correcting software. Built on the canonical ASH 9-bit state model, it provides deterministic diagnostics, recovery, fallback, containment, and safe-failure semantics — with native implementations for Windows, macOS, and iOS.
- Deterministic diagnostics, recovery, and safe failure
- Canonical 9-bit ASH state space (512 states, fixed codeword set)
- Native C++20 edition for Windows (x64 & ARM64)
- Apple-native Swift package for macOS & iOS
- Conformance-tested against a fixed canonical baseline
Galdr
Galdr is an ASH-based bitwise compression engine for compact AI. Built around syndrome algebra, adinkra transforms, XOR repair, and hypercube state encoding, Galdr compresses neural-network structures into fast, locally repairable binary representations. Designed for Ennea AI, it enables smaller, more efficient models that scale through recursive structure instead of massive floating-point computation.
- ASH-based bitwise compression engine for compact AI
- Syndrome algebra and XOR repair for local recovery
- Adinkra transforms and hypercube state encoding methods
- Fast, locally repairable binary representations of neural networks
- Recursive structure instead of massive floating-point computation
- Smaller, more efficient models designed for Ennea AI
Ennea AI
Ennea AI is a new kind of neural-network architecture built on the ASH Model: a recursive 9-dimensional hypercube where every vertex can unfold into another 9D Enneract. Instead of relying on massive floating-point weight matrices, Ennea uses Galdr bitwise compression to encode, repair, and route intelligence through compact binary structures. The result is an AI foundation designed to be highly scalable, fast, energy-efficient, and radically smaller than conventional neural networks.
- New neural-network architecture built on the ASH Model
- Recursive 9-dimensional Enneract hypercube structure
- Every vertex can unfold into another 9D Enneract
- Galdr bitwise compression to encode, repair, and route intelligence
- No massive floating-point weight matrices
- Highly scalable, fast, energy-efficient, and radically smaller
Built Different, On Purpose
Every one of my projects follows the same principles: architectural clarity, modular integrity, and the discipline to prioritize long-term reliability over short-term convenience.
No Black Boxes
Every component has clear inputs, outputs, and boundaries. Systems are debuggable, auditable, and transparent by design.
Standards-First
I build on proven standards — C++20, modern platform SDKs, and established architectural patterns — not trends.
Built for Clean Handoff
Projects are documented and structured with clean boundaries so they can be maintained, demonstrated, published, or licensed deliberately when appropriate.
Interested in the Architecture?
Questions about the architecture, research direction, applications, or focused collaboration are welcome.
Contact Me