// Active Research · Tools · Experiments

The
Lab

Active Dev
Deployed
Archived

The question that has always driven the work: how does music live inside an experience that's different every time? The Lab is where that question gets stress-tested, broken, and rebuilt.

These are the active systems, deployed tools, and documented experiments built at the intersection of game audio, artificial intelligence, and adaptive music theory. Some are infrastructure. Some are instruments. Some are just questions that needed code to answer them.

Active Systems
Active Development
Investigative
Templater
Four-layer MCP system for natural language control of Wwise

The most technically ambitious project in the lab. A four-layer Model Context Protocol architecture that bridges natural language commands with Wwise audio middleware via WAAPI — describe what you want in plain English, and the system interprets, translates, and executes it in the game engine.

The workflow that should have always existed. Once you've spent time in both the creative chair and the implementation chair, it's the most obvious thing imaginable.

Built on a deep understanding of how composers actually think about interactive music — not how engineers document it. The system speaks the language of creative intent.

MCPWAAPIWwise Claude APINode.jsAdaptive Music
Architecture · Four Layers
01
Natural Language InterfaceInterprets creative intent in plain English. Maps vocabulary to Wwise architecture patterns.
Claude API
02
MCP Server Core43 music-focused tools for batch operations, transition editing, and state management.
MCP / Node
03
WAAPI BridgeDirect WebSocket connection to Wwise Authoring API. Bidirectional, real-time.
WAAPI / WS
04
Wwise EngineThe target. Every layer exists to make this one behave exactly as imagined.
Wwise
● Active Research · Methodology
GMA Standard

A compositional philosophy and prompt methodology developed over 13+ months of documented experimentation with Suno AI. Approaches AI music generation as a music director, not a consumer. Literary vocabulary over genre tags. Raw vulnerability over coolness. Leave space for the unexpected.

Suno AIPrompt EngineeringMusic Theory
● Active Generative Engine · Prototype
DRIFT

Generative ambient soundscape engine for real-time interactive environments. Synthesizes rather than plays back — so it never repeats, never anticipates, and never competes with the action. Web Audio API, four distinct modes based on musical intervals.

Web Audio APIGenerativeReal-time Synthesis
// Note
On building things before the field catches up
Every tool in this lab started with the same feeling: this should already exist. The Investigative Templater exists because describing what you want in plain language and having the system build it is the most obvious workflow imaginable — once you've spent enough time in both the creative chair and the implementation chair. The GMA Standard exists because AI music generation tools were being approached as search engines rather than instruments. DRIFT exists because "adaptive ambient" is still mostly a toggle switch in a game engine property sheet.
Deployed Tools
GMA-assist Prompt GeneratorSuno AI prompt structuring with dynamic descriptor output. Clean + bracketed versions. GMA Standard compliant.
Suno AI
Prompt Engine
Open →
System Interrupts50+ oblique creative prompts for POUROVER sessions. Questions that cut through avoidance. Press space.
JavaScript
Generative Text
Open →
Improv Jam Spark GeneratorRandom constraints for live sessions. Key, groove, vibe, keyboard sound, tempo, who starts, one forbidden element.
JavaScript
Music Theory
Open →
Numbers Station K-404Chord progression encoder/decoder. Encode messages as playable progressions. Transmit via Tone.js.
Tone.js
Music / Cipher
Open →
Tempo TapperBPM training game with 40+ song database and leaderboard. Real-time metronome feedback. Tap or spacebar.
Tone.js
Game / Training
Open →
Sweetened TunerCustom tuning reference based on James Taylor's personal intonation preferences. Equal temperament is a compromise.
Web Audio API
Tuning / Acoustics
Open →
Experiments & Archive

Built to learn. Shipped anyway. Some became methodologies, some became tools, some just needed to exist long enough to reveal the question underneath them.

Archived · Concept
Audio Interface // Manifesto
A typewriter-on-paper blog about AI and audio. "The Universal API is Sight." Correct thesis. Wrong format. One post published.
Archived · Campaign Tool
Kinetic Protocol
Full campaign execution dashboard for POUROVER's Kinetic Syntax release. Three-day social rollout, verb simulator, email templates. "We are not nouns. We are verbs."
Archived · Proposal
IEAI
Investor-facing proposal for an Interactive Entertainment Audio Institute in Traverse City. Full program structure, faculty model. Still believe in this. Waiting for the right partner.
Archived · Satire
Corporate Apology Generator 3000
Generative DEI-buzzword-laden corporate apologies from three inputs. PDF export. Built in one afternoon after reading one too many press releases. Accurate enough to be concerning.
Active · Service
Source-First Method
Consulting and training for musicians approaching AI production. $100/hr. You can't prompt a tool you don't understand. Start with your own sound.
Archived · Experiment
Code + Cadence
Generative visuals tied to harmonic data, algorithmic notation experiments, and browser-based instruments that shouldn't work but do.