Analyse a song on Audial and press Get audio features: you get its energy and danceability (0 to 100% on the website, 0 to 1 in the API) with its loudness, key and mode, and over the API, the Python SDK and MCP the same measures for every section, so a drop and its breakdown can be compared.
$10 a month or $100 a year. Unlimited runs, cancel any time. 2 free runs to try it first.
Audio features put numbers on how a song feels. Analyse a track, press Get audio features, and you get its energy, danceability, loudness, key, mode and beat division. Over the API, the Python SDK and MCP, every section of the analysis also gets its own energy, danceability, loudness, mode and beat division, lined up with the section's start and end.
Energy reflects loudness, rhythmic activity and low end. Danceability reflects how strong and regular the beat is. Because both are measured per section, a drop reads higher than its breakdown, and the parts of a set or a playlist can be ordered by how hard they hit. Loudness is integrated loudness in LUFS, measured to the broadcast standard over both channels. Mode is one of nine (major, natural, harmonic and melodic minor, dorian, mixolydian, lydian, phrygian and locrian) with a confidence, and beat division is the smallest regular subdivision, from 1/4 to 1/32.
Audio features read the mix, so they start as soon as the analysis has finished, with no wait for the track to be prepared. The website shows the song's values in its Musical Properties and Technical Details tiles. Audio features are included in the analysis's run.
What you get
- Energy and danceability for the song, and over the API for every section
- Loudness in LUFS, measured to the broadcast standard over both channels
- The song's key, and the mode of the song and each section, with a confidence
- Beat division: the smallest regular subdivision, from 1/4 to 1/32
- Starts as soon as the analysis finishes, because it reads only the mix
- Included in the analysis's run; on the website, by REST, in Python and through MCP
Hear it
- Lead-in
- Intro
- Up
- Chorus
- Up
- Chorus
- Down
- Outro
This mix's sections, from its analysis. Get audio features measures energy, danceability, loudness, mode and beat division for each section, and for the song as a whole.
Audial vs doing it by hand
| Task | Audial | By hand |
|---|---|---|
| Loudness | Audial: Integrated LUFS for the song and for each section | By hand: A loudness meter in your DAW, played through each section; accurate, and free if you own one |
| Energy and danceability | Audial: A number per section, measured the same way every time | By hand: A judgement by ear: fine for one track, hard to keep consistent across a crate |
| Key and mode | Audial: Detected, with a confidence for the mode | By hand: By ear at a keyboard; modes beyond major and minor take a trained ear |
| Per section | Audial: Lined up with the analysis's sections | By hand: Mark the sections first, then judge or meter each one |
| Whether it works on the night | Audial: Measures the audio, not the room | By hand: You know whether a track works at 2 a.m.; no number does |
Use it from code
# 1. Analyse the mix. Keep the exeId from the answer.
curl -X POST "https://api.audialmusic.ai/api/functions/run/segmentation" \
-H "X-API-Key: your_api_key" -H "X-User-ID: your_user_id" \
-F "userId=your_user_id" \
-F "original=@/path/to/track.mp3"
# 2. Audio features on that analysis: they start as soon as it has finished.
curl -X POST "https://api.audialmusic.ai/api/functions/run/audio-features" \
-H "X-API-Key: your_api_key" -H "X-User-ID: your_user_id" \
-F "userId=your_user_id" \
-F "analysisExeId=exe_12345abcde"
# 3. The song's values are in song_analysis, each section's in segment_analysis[i].
curl "https://api.audialmusic.ai/api/db/your_user_id/execution/exe_12345abcde/analysis" \
-H "X-API-Key: your_api_key" -H "X-User-ID: your_user_id"import audial # audial-sdk 1.7 or later
audial.config.set_api_key("your_api_key")
audial.config.set_user_id("your_user_id")
analysis = audial.segment("track.mp3")
features = audial.audio_features(analysis["analysis_id"])
song = features["song"]
print(song["key"], song["mode"], song["loudness"], "LUFS")
for s in features["sections"]: # 0-1; None when a value could not be measured
print(s["label"], s["start"], s["end"], s["energy"], s["danceability"])# In Claude, Cursor or any MCP client with audial-mcp 0.6.0 or later installed:
"Which section of ~/Music/track.mp3 has the most energy, and how danceable is it?"Frequently asked questions
Upload it on the Audio Analysis tab and press Get audio features: the Musical Properties tile shows the song's energy and danceability as percentages. For each section's values, call /functions/run/audio-features with the analysis's exeId, or audial.audio_features in Python.
Energy reflects loudness, rhythmic activity and low end. Danceability reflects how strong and regular the beat is. Both are measured per section, so a drop reads higher than its breakdown.
Because a song is not one level. An intro, a breakdown and a drop differ, and one average for the whole track hides that. Values per section show where a song lifts and where it rests, which helps when you pick the part to sample, cut an edit or plan a set.
Energy and danceability run from 0 to 1 in the API and show as 0 to 100% on the website. Loudness is integrated LUFS, a negative number that is louder the closer it is to 0. Mode comes with a confidence from 0 to 1.
Only loosely. Every service defines energy and danceability its own way, so compare tracks Audial measured with each other rather than mixing sources. Audial's own values were rebuilt in November 2026 and now differ between tracks and sections, so an analysis from before then reads differently.
Yes. Post { userId, analysisExeId } to https://api.audialmusic.ai/api/functions/run/audio-features with your X-API-Key and X-User-ID headers. The answer comes at once with state queued, and the result lands on the analysis. Calling again answers the same run; rerun: true starts a new one.