How Files Are Organized

When you drop files into Producer Dashboard’s inbox folders, the app automatically figures out which song each file belongs to and organizes everything into a clean folder structure. This page explains why the inbox folders are designed the way they are, what happens step-by-step during import, and what you can configure in Settings.

---

Why There Are Separate Inbox Folders

Producer Dashboard uses three distinct inbox folders:

• Bounces/ — for exported audio mixes
• Stems/ — for individual instrument or vocal tracks
• Projects/ — for DAW session files

The reason for separate folders, not one: bounces and stems use exactly the same file extensions (.wav, .mp3, .aiff, etc.). There is no way to tell them apart by looking at the file itself — a stem and a bounce are both just audio files. The folder you put them in is the only signal the app has to classify them correctly.

If both types went into a single folder, every stem would end up categorized as a bounce (or vice versa), which would break how tracks are tagged, how the waveform panel separates them, and how they’re displayed in the activity panel.

Projects have distinct extensions (.als, .logicx, .flp, etc.) so they could technically share a folder — but keeping them separate makes your inbox easier to manage and keeps the pipeline logic clean.

Note

Case-insensitive: The app recognises bounces/, Bounces/, and BOUNCES/ as the same folder.

---

What Happens When You Add Files

As soon as files land in the inbox folders and you trigger an import (from the Files page), the app runs them through a four-stage pipeline automatically.

Stage 1 — Scan

The app scans all three inbox folders and classifies every file by extension and location:

Inbox Folder	What goes here
Bounces/	WAV, MP3, AIFF, FLAC, M4A
Stems/	WAV, MP3, AIFF, FLAC, M4A
Projects/	.als, .logicx, .flp, .ptx, .cpr, .rpp, .reason, .song

Hidden files (names starting with .) are skipped. Files already linked to an existing track group are also skipped so they aren’t processed again.

Ableton Live projects: when the scanner encounters a .als file, it treats the entire parent folder as the unit to move — not just the file — because Ableton projects depend on a Samples/ subfolder and other assets alongside the .als.

Stage 2 — Group

This is where the app decides which files belong to the same song. It runs a two-phase matching algorithm (detailed below).

Stage 3 — Organize

Matched files are moved from the inbox into a structured tracks/ hierarchy:

tracks/
├── Ancient Dusk/
│   ├── bounces/
│   │   └── Ancient Dusk - Clean.wav
│   ├── stems/
│   │   ├── Ancient Dusk Drums.wav
│   │   └── Ancient Dusk Bass.wav
│   └── projects/
│       └── Ancient Dusk Project.als
└── Midnight City/
    ├── bounces/
    └── stems/

For project files that already exist in the organized hierarchy, the inbox copy is replaced with a symbolic link rather than duplicated. This keeps your projects folder clean and saves disk space.

Inbox files that were already in the tracks/ folder are cleaned up automatically — you won’t end up with duplicates.

Stems that don’t match any song stay in the Stems/ inbox rather than being moved. Only bounces and projects create new track groups; unmatched stems wait until a matching group exists.

Stage 4 — Sync

Organized track groups and their files are synced to the cloud database, making them available across devices and visible in your dashboard grid.

---

The Grouping Algorithm

Getting files into the right song group is the hardest part of import. The app uses a two-phase approach.

Filename normalization

Before anything is compared, every filename is stripped down to just the core song name. The normalizer removes:

• File extension
• Underscores (replaced with spaces)
• Version numbers: v1, V3, version 2
• Timestamps: 2.06.22 am, [2025-10-09], 6-digit dates
• Content in parentheses or square brackets: (final), [127 D#m]
• Mix identifiers: clean, full mix, radio edit, demo, wip, master, mastered, rough, final (planned — not yet implemented, see note below)
• Technical specs: 320kbps, 44.1hz, 128bpm
• Trailing punctuation and special characters

Examples:

Filename	Normalized to
ATL - Ancient Dusk [127 D#m] v11 - CLEAN.wav	atl ancient dusk
Ancient Dusk - Full Mix.wav	ancient dusk
Ancient Dusk Radio Edit.mp3	ancient dusk
My Song Beat_Idea_v3_WIP.wav	my song beat idea
Clean Slate.wav	clean slate (“clean” only stripped as a trailing suffix, not mid-name)

Note

Mix identifier stripping (clean, full mix, demo, wip, etc.) is a planned feature (PRO-56) and is not yet implemented. Currently, these words are not stripped during normalization, which means Ancient Dusk - Clean.wav and Ancient Dusk - Full Mix.wav may not automatically group together unless the core song name matches exactly via Phase 1 or scores above 0.85 in Phase 2.

When implemented, the normalizer will be careful not to strip words that are part of the actual song title — Clean Slate would stay clean slate because clean isn’t in trailing position.

Phase 1 — Match against existing library

Every file is checked against your existing track groups in the database. The app looks for your existing song names as a substring within each normalized filename, checking longest names first so the most specific match wins.

Existing groups: "Atl Alex Ancient Dusk", "Atl Alex"

Incoming files:
  "atl alex ancient dusk clean"    → "atl alex ancient dusk" found → matched ✓
  "atl alex ancient dusk full mix" → "atl alex ancient dusk" found → matched ✓
  "midnight bloom v3"              → no match → goes to Phase 2
  "midnight bloom final mix"       → no match → goes to Phase 2

Matched files are assigned to the existing group immediately. They never create duplicates.

Phase 2 — Cluster new files together

Files that didn’t match anything in your library are clustered against each other using Jaro-Winkler string similarity. Any two files whose normalized names score 0.85 or higher get grouped together.

The app is also noise-aware: if a large portion of the unmatched filenames share a common leading word (an artist prefix like “Atl Alex” or “Es” that appears in 40%+ of files), that token is stripped for comparison purposes only, so it doesn’t inflate similarity scores between unrelated songs.

Unmatched files:
  "midnight bloom v3"
  "midnight bloom final mix"
  "midnight bloom bass stem"
  "random idea"

Similarity check:
  "midnight bloom" ↔ "midnight bloom" → score ≥ 0.85 → same cluster
  "midnight bloom" ↔ "random idea"    → score < 0.85 → different cluster

Result:
  Cluster 1 → group name: "Midnight Bloom" (shortest name in cluster)
  Cluster 2 → group name: "Random Idea"

The group name is taken from the shortest normalized name in each cluster, then title-cased. This tends to produce the cleanest result since shorter names have had more suffixes stripped.

---

Settings: Storage Folder

You can view and manage your storage folder from Settings → Storage.

Dropbox mode

If your account is connected to Dropbox, the storage folder path is managed by Dropbox and is read-only. The inbox folders (Bounces/, Stems/, Projects/) live inside your PRODUCER-DASHBOARD folder on Dropbox.

Local mode

If you’re not using Dropbox, the default storage folder is ~/Music/ProducerDashboard. You can change this to any folder on your computer:

1. Go to Settings → Storage
2. Click Change next to the current path
3. Choose a folder using the system directory picker

The app validates that the chosen path exists, creates the Bounces/, Stems/, Projects/, and Reports/ subfolders if they’re missing, and saves the path so it persists across app restarts.

You can also click Open Folder to reveal the storage directory in Finder (or File Explorer on Windows) at any time.

Note

The Change button is only available in local mode. If you’re connected to Dropbox, the path is controlled by Dropbox and cannot be changed here.

---

Tips

• Drop stems in Stems/ only — there’s no other way for the app to know a file is a stem. Any audio file in Bounces/ is treated as a bounce regardless of its name.

• Name files with the song name first — the normalizer strips suffixes from the end. Ancient Dusk - Clean.wav groups correctly; CLEAN_AncientDusk_v3.wav may not.

• Import regularly rather than in bulk — the more songs already in your library, the better Phase 1 matching works. Importing incrementally reduces the chance of files clustering incorrectly in Phase 2.

• Unmatched stems stay in the inbox — if you see stem files still in Stems/ after an import, it means no matching song group was found. Import the corresponding bounce or project first, then import the stems again.

---

Related

• Connecting Dropbox — sync files across devices
• Managing Dropbox Connection — troubleshoot sync issues
• Importing Tracks — trigger an import from the Files page
• Notifications — stay updated on import completions