MIDI Composer UI (Next.js + Postgres + OpenAI)

This is a Next.js app that provides a UI and API for:

• Uploading a MIDI file
• Optionally providing genre + subgenre
• Selecting additional instruments to be AI-composed (each instrument = track)
• Optionally defining song sections with timestamps and per-section instrument selection
• Pressing Compose to run an AI-powered composer and download the resulting MIDI

✅ The server-side composer uses OpenAI to generate musical arrangements that layer on top of your MIDI input.

Tech

• Next.js App Router (React)
• Postgres via Prisma
• OpenAI API for AI music composition
• @tonejs/midi for MIDI parsing and generation

Local setup

1. Install dependencies:

npm install

2. Create a Postgres DB and set environment variables:

cp .env.example .env

Edit .env and set:

• DATABASE_URL - your Postgres connection string
• OPENAI_API_KEY - your OpenAI API key for AI composition

3. Generate Prisma client and run migrations:

npm run prisma:generate
npm run prisma:migrate

4. Start the dev server:

npm run dev

Open http://localhost:3000

API endpoints

• POST /api/jobs – accepts multipart/form-data with:
  • midiFile (File)
  • genre (string)
  • subgenre (string)
  • instruments (JSON array of instrument ids)
  • sections (JSON array of section configs)
• POST /api/jobs/:jobId/compose – runs the AI composer to generate musical arrangements
• GET /api/jobs/:jobId – job status + metadata
• GET /api/jobs/:jobId/download – streams the output MIDI

Where files go

• Original uploads: ./uploads
• Output files: ./outputs

In production/serverless, you'd likely move these to object storage (S3/R2/GCS) and store URLs in Postgres.

Next iteration ideas

• Background job queue (BullMQ / pg-boss) + progress polling for long-running compositions
• Authentication + user libraries
• Fine-tune composition parameters (temperature, creativity level)
• Support for more AI models (Claude, custom music generation models)
• Real-time composition preview