Producers
Use it as an audio analyzer for reference tracks before you mix: loudness, tempo, key and how the sections are built, next to what you hear in the arrangement.
Drop an MP3, WAV, M4A, FLAC, OGG or AAC file of up to 10 minutes into the song analyzer: a finished master, a rough demo, or a reference track you are learning from.
An AI listening model describes the genre, mood, instruments, vocals and sections, while an audio analyzer measures tempo, key, tuning, loudness and chords from the file itself. It usually takes a minute or two.
Start with the one-line profile, then go through the structure timeline, the measurements and the chords. Play any section straight from the report.
The song analyzer runs two kinds of analysis on one upload: how the song sounds to a listener, and what a meter reads from the audio.
The main genre, the closest subgenres and style tags, the mood in a few words, and how the energy moves from start to finish.
One sentence that places the sound in an era, a scene or a production tradition, written without naming artists.
The instruments you can hear, most prominent first, plus the voice type, delivery and language of the vocals.
Intro, verse, chorus, bridge and the rest, each with start and end times and a note on what changes there.
Tempo with half-time and double-time readings, the key with its Camelot code and compatible keys, and the A4 reference pitch.
Integrated LUFS, loudness range and sample peak for mix and master checks, plus the most used chords with Roman numerals.
This is the report for The Last Long Goodbye, a piano pop ballad made on Lacuna. Every value comes from the real analysis of the song, nothing is edited by hand. (2:39)
Detected as major and minor triads only, so sevenths, suspensions and slash chords show up as their nearest triad.
Genre labels on streaming services describe an artist, not a track. This genre finder listens to the song you upload and names what it actually sounds like.
You get the broad genre a listener would use, like country, hip hop or electronic, and the narrower tags that place it, like honky-tonk ballad, boom bap or melodic techno.
Unlike a genre finder that stops at a label, the report also lists the instruments, drum feel, vocal delivery and production notes that put the song in that genre, so you can check the call yourself.
Every report ends with a style description you can carry into a new original song on the AI Song Generator, written without artist names.
Same song analyzer report, different reasons to read it.
Use it as an audio analyzer for reference tracks before you mix: loudness, tempo, key and how the sections are built, next to what you hear in the arrangement.
Get the key, the tempo and the chord loop before you start working a song out by ear, and see where each section begins.
Writing up a song analysis for class? Start from the form, the instrumentation and the harmony, then put it in your own words.
A music genre identifier for anyone who wonders what genre a song is: the answer comes from the sound itself, with the details that make it that genre.
Sign in, drop the song into the song analyzer above and select Analyze song. The report appears in your list, usually within a minute or two, and stays there so you can come back to it.
Yes, it works as a song genre finder. The report names the main genre and the closest subgenres from the audio itself, not from the artist's catalog page. It also lists the instruments, vocals and production details behind that call, so the label comes with its reasons.
They come from the AI listening model in this music analyzer, so treat them like a trained listener's notes: usually right on clear-cut songs, less certain on genre blends. Anything the model couldn't tell from the audio is listed under Hard to tell from the audio instead of being guessed.
An audio analyzer measures them with signal analysis on the whole song, not the listening model. Tempo comes from beat tracking and key from the pitch content across the track. For a quick check of many files without uploading anything, the BPM & Key Finder runs in your browser.
Beat trackers often lock onto twice or half the pulse you feel, especially in ballads, trap and drum and bass. A 140 BPM reading on a slow country ballad usually means the felt tempo is 70, so both readings are listed next to the main one.
LUFS is integrated loudness over the whole song; streaming services turn songs down to around −14 LUFS. Loudness range (LRA) shows how much the level moves between quiet and loud parts. Sample peak is the highest level in the file, in dBFS, measured on a mono mix.
Camelot is DJ notation for keys: a number from 1 to 12 and a letter, A for minor and B for major. Keys with the same number, one step away, or the same number on the other letter mix without clashing. The report lists the compatible codes for your song.
No, they're approximate. Chord detection reads major and minor triads only, so a G7 or a Dsus4 shows up as its nearest triad. The most used chords and the opening progression are a reliable starting point for working the song out by ear.
The measurements and chords cover the whole file, up to 10 minutes. The listening notes cover the first 4 minutes; for a longer song, the report says exactly which part was heard.
Each song costs 10 credits, and a free account comes with a monthly credit allowance, so you can start without paying. If the analysis fails, the credits come back automatically. If only the listening notes fail, the measurements are still delivered and 6 of the 10 credits come back.
Uploaded songs and reports are tied to your account. They aren't shown to other users or added to any public page, and you can delete a report from your list at any time.
Use the style description to write a new original song on the AI Song Generator, move the song to another key with the Song Key Changer, or pull out the vocals and instruments with the Stem Splitter.
Upload a track to the song analyzer and get genre, mood, structure, tempo, key, loudness and chords in one report.