Music Energy Analyzer

Local audio analysis

Rank your music by how it feels

Every track is analysed in your browser: intensity from 0 (ambient) to 100 (paroxysmal), tempo, key, mood, structure, genre. Then sort, compare and build playlists that follow the curve you want.

Read the guide

Where to start?

I have my audio files

Drop them (or a whole folder) anywhere on the page, or use Import.
Library: scores, curves, key, mood; open a track to see everything.
Tune it to how you feel: lyrics, duels, genres, games.
Set: draw a curve, get an ordered playlist.

I have a Spotify playlist

Spotify: log in (once) and import the playlist.
Live: the app plays each track on your Spotify and analyses it in real time.
Map, comparison, split, sorted playlist created on your account.

I want to see first

Test bench: synthetic tracks that move a single variable, from min to max.
Live › Demo mode: the live scan without Spotify.
Test tracks also let you try Set, duels, games and genres.

The tabs

Your files never leave your browser: only the measures are kept (IndexedDB). The score is a practical perceptual ranking tool, not a scientific measure.

Spotify playlistbuild its progression, add or remove tracks, sort it in place

Playlist

With a playlist selected, “Build a progression” orders only its tracks; ✕ removes a track from the playlist. Sorting only moves tracks: analysed ones take the order above, the others stay at the end, and the added dates are kept. Only your playlists and the collaborative ones can be changed.

Click “Build a progression” to order the analysed tracks.

Target curve

Drag the points to draw the intensity you want over the set · click: add a point · double-click: remove it. The bars show the generated set.

Tracks Target length min (0 = every track) First track Last track Never the same artist twice in a row

Weight of each criterion

Split into several playlists

Splitting settings — if the detection is off

Instruments (lanes) Sensitivity Weak note filter Sharing between instruments Min. gap between notes (ms) Silence gate (dB) Resolution (bands / octave) Min. frequency (Hz) Max. frequency (Hz)

The last three settings rerun the spectral analysis (a few seconds); the others apply instantly.

0:00.0 Play Cues Music Cues Zoom 10 s

Tick the lanes that make the map: they sound as cues and count in the KPS and difficulty. Click the matrix or a curve: play from there, click again: stop. Space: play / stop. Wheel: scroll, Ctrl + wheel: zoom.

computed on the ticked lanes

Sort a Spotify playlist

Import the track list of one of your playlists, match it with your local audio files (analysed here) or scan it live, then create on your account a new playlist ordered by intensity. Spotify does not provide the audio: only what you play is analysed. Your original playlist is never changed.

On developer.spotify.com/dashboard, create an app (the owner account must be Premium). Tick Web API.
Add this address in Redirect URIs :
In User Management, add your Spotify account's e-mail address (development mode: 5 accounts max).
Paste here the Client ID (public; never use the Client secret) :

No files? : the Live tab plays each track on your Spotify and analyses it in real time.

You have the files: drop them anywhere on the page (or use Import), they are matched automatically (ISRC, then title / artist / length). Fix a match with the row's menu.

#	Spotify track	Length	Analysis source	Score

Genres Genres come from MusicBrainz (track, album, artist) and are looked up automatically. They drive the genre filter, grouping by style and suggestions (you can always fix them).

Playlist map Compare with

Each dot is an analysed track: intensity across, mood (dark → bright) up. Click: details.

New ordered playlist

Order Tracks not analysed Name

Tracks and metadata provided by Spotify. This app is neither affiliated with nor endorsed by Spotify.

Test bench

Synthetic tracks, made in your browser, that move a single variable at a time (min, mid, max) while the rest stays put, plus a few sweeps where a variable changes over time. They go through exactly the same analysis as your files. The grid compares what the algorithm measures with what is expected.

They are caricatures: they check that every measure reacts in the right direction and over its whole range, not that the score is right on real music. Test tracks join the library, tagged “Test”, and also let you try the Set, duels, games, splitting or the Live tab's demo mode.

Scan settings

1 · Sound capture System audio (screen sharing: “Entire screen” + “Also share system audio”) Audio input (VB-Cable: silent scan)

off

2 · Spotify

Import a playlist in the Spotify tab first.

Demo mode: without Spotify or audio sharing, with synthetic tracks from the test bench

3 · Analysis mode

Budget s / track Max gap between probes s

Skip tracks already analysed from a file Re-analyse tracks already captured Only re-analyse tracks analysed with an older version

Before starting (Windows)

Spotify › Settings › Playback: turn off Normalize volume (the “Loud” mode compresses the dynamics) and crossfade.
System audio: nothing else must play during the scan (videos, notifications). The sound must stay audible.
Silent scan: install VB-Cable, then Windows Settings › Sound › Volume mixer › Spotify › Output = “CABLE Input”. Here, pick the “CABLE Output” input.
The audio is never recorded: it is analysed in memory then discarded, only the measures are kept.

Ready.

♪

Waiting

No track

Start a scan to analyse your playlist live.

0:000:00

Intensity curve

Profilecurrent window · track

Spectrum

LoudnessLUFS

Live measures

Scan queue

Play order ▶ on a track: analyse it now

Session

Captured tracks join the library (sorting, curves, progression, playback) and the Spotify tab's sorted playlist.