Selected work / 07

ProgressionLab

An AI-assisted harmony and songwriting workspace that turns musical intent into playable chord progressions, practical voicings, visual explanations, and exportable session material.

My role: Creator and independent product engineer
Scope: Product strategy, interaction design, full-stack engineering, AI orchestration, and music playback
Stage: Deployed public web product
Core technologies: Next.js · TypeScript · OpenAI Responses API · Tone.js

ProgressionLab generator with seed chord, mood, mode, genre, style reference, and adventurousness controls

Case study

Turning harmonic intent into something a musician can hear, inspect, and keep.

The problem

Music ideas lose momentum when harmonic exploration, playable voicings, audio preview, and export are split across disconnected tools. A generated answer is not enough; the material has to fit a musician’s working process.

What I built

I designed and built a deployed browser product that joins structured AI generation, music-theory logic, tempo-aware playback, piano and guitar diagrams, progression management, public sharing, and MIDI and PDF export. I also built a separate operational surface for prompts, marketing content, translations, analytics, entitlements, and security controls.

Key decisions

01Constrain model output
Strict JSON schemas and deterministic voicing normalization turn generated responses into typed chord, progression, and arrangement data the interface can reliably use.
02Keep music interactive
Tone.js playback, visual voicings, diagrams, and MIDI-aware exploration keep the experience grounded in sound and performance rather than text alone.
03Build the operating controls too
Prompt versioning, entitlements, marketing draft and rollback workflows, human-reviewed translation, analytics, WebAuthn, and a separate admin deployment make the product operable beyond one demo flow.

Inside the work / 03 views

A complete loop for harmonic exploration

The product connects musical intent to structured suggestions, playable voicings, audible previews, reusable examples, public sharing, and session-ready exports.

ProgressionLab result showing a chord suggestion, musical rationale, playback, MIDI and PDF controls, voicing notes, and a piano diagram — 02
**Playable suggestion**
A generated idea becomes inspectable product data with rationale, tension and confidence cues, practical voicings, playback, and export controls.

ProgressionLab public examples library with filters and curated beginner chord progressions — 03
**Examples library**
Curated starting points and filters help musicians continue from a useful progression instead of beginning with an empty prompt every time.

Technical view

Architecture

Playable AI music workspace12 components · 14 connections

Musician-facing tools share one typed application boundary, while persistence, prompts, access policy, and operational controls remain behind server-owned interfaces.

Creative surfacesHear, inspect, and continue ideas

Application + music logicTyped generation and playback boundary

Persistence + operationsAccess, content, prompts, and telemetry

Next.js · React
Harmony generator
Intent, mood, mode, and style
Web MIDI · Canvas
Chord Explorer
MIDI input and harmonic context
React · App Router
Library + sharing
Saved, public, and curated ideas
Next.js route handlers
Generation API
Validation, access, and normalization
OpenAI · strict schema
Structured suggestions
Chords, voicings, and structure
Tone.js
Playback engine
Tempo-aware chord audition
MIDI · PDF
Session exports
Portable musical material
Typed TypeScript
Music-theory model
Voicings, diagrams, and inference
Prisma · PostgreSQL
Product record
Users, progressions, and prompts
Separate Next.js app
Operations console
Prompts, content, pricing, analytics
Sessions · WebAuthn
Identity + entitlement
Protected actions and security keys
Sentry · server events
Operational telemetry
Errors, usage, and funnel events

surfaceservicedataaiintegrationcontrolruntime

Read system connections

Harmony generator → Generation API
Chord Explorer → Music-theory model
Library + sharing → Generation API
Generation API → Structured suggestions
Generation API → Music-theory model
Music-theory model → Playback engine
Music-theory model → Session exports
Generation API → Product record
Generation API → Identity + entitlement
Operations console → Product record
Operations console → Structured suggestionsprompts
Library + sharing → Product record
Generation API → Operational telemetry
Operations console → Operational telemetry

Built so far

Turns normalized musical inputs into typed chord, voicing, progression, and song-structure suggestions using strict structured output and versioned prompt fallback behavior.
Connects browser playback and piano or guitar diagrams to saved progressions, public sharing, and MIDI and PDF export.
Separates musician-facing workflows from admin controls for prompt versions, marketing content, translation drafts, pricing configuration, and funnel analytics.

Where it stands

The public product is deployed, but user count, revenue, retention, conversion, and production-scale usage have not been independently established.
AI suggestions support musical exploration; they do not establish musical correctness or replace the musician’s judgment.

Repository Live demo

Continue exploring

Next case study

View all selected work

Geospatial / SAR evidence systems

EchoAtlas

Public portfolio demonstration

Open case study