Frequently asked questions
What does Attracr do?
Attracr analyses the music in a public Spotify playlist and shows you the sound it adds up to: which tracks sit at the centre of it, which sit at the edge, and how the playlist's followers have moved over time. It suggests tracks that fit. You decide what to do about them.
Can Attracr change or delete my playlists?
No. Attracr cannot change your playlists — not by policy, by architecture. No code in Attracr writes to Spotify. Every suggestion is applied by you, in Spotify. That is more work for you, and it is also the reason nothing on your account can be damaged by a bug on ours.
What does signing in with Spotify give Attracr access to?
Your Spotify or Google login identifies you, and nothing more. The token is never used to read your library. The Spotify permission Attracr asks for is your email address; there is no permission requested for playlists, saved tracks, or listening history.
Where does the data come from?
Attracr reads the same public Spotify pages anyone can open in a browser. Follower counts, track listings, album and artist information — all of it is visible without logging in. Attracr reads no private playlists.
Does Harmony use the mean or the median?
Harmony compares each scored feature with the playlist's median: the middle value after sorting, or the average of the two middle values for an even count. The arithmetic mean is what people usually call the average. A few extreme tracks pull the mean more than the median. The radar chart's Playlist Average uses the mean; Harmony's tolerance bars use medians, with half-time and double-time correction for tempo.
Energy, loudness, mood, speechiness and tempo contribute to the score. Danceability and production style use majority checks instead; their flags do not change the score. Missing measurements do not set the centre, so analysis can change as more tracks are measured. A partially measured track can shift the bands of the features it has but earns no score credit. The score needs at least three fully measured tracks, and counts as reliable when at least 60% of the playlist's entries are scored.
What happens to Harmony when I add or remove tracks?
After Attracr detects an edit in Spotify, it recalculates the playlist's centre and reassesses all measured tracks, including the ones you kept. A track can become an outlier without changing itself: the playlist around it changed. Removing outliers does not guarantee a higher score, and a large edit can move the median sharply. Core tracks currently protect against removal suggestions; they do not hold the centre in place.
Scheduled playlist checks run daily. A manual refresh requests analysis after a changed playlist is read; if that request is missed, an hourly job discovers the change and catches up. Reading the playlist and analysing it are separate steps: a finished refresh does not mean the new Harmony or candidates are ready, or that audio analysis of new tracks has finished. Added tracks count once their audio is analysed.
While an analysis is pending, the playlist page says Analysis updating and keeps showing the last completed analysis. The page checks for about three minutes after you open it or refresh. If the analysis takes longer, it says so and offers Check again; the analysis carries on either way.
How does Attracr suggest tracks to add?
Every tracked playlist has a Candidates tab. Attracr listens to the tracks already in it, works out the sound they share, and searches its catalogue for tracks that sound the same and are not in the playlist yet. One list combines Style and Mood suggestions with tracks from your buffer. These searches use learned descriptions of sound, not just Harmony's medians. They can follow several distinct sounds within a playlist. A Within range label checks energy, mood, loudness and speechiness; it does not guarantee a track will pass every Harmony check after you add it.
Do candidates change when my playlist changes?
When the analysis changes, Attracr searches again against the playlist's current sounds, and the catalogue those searches draw from refreshes daily. Until a new search finishes, the tab keeps the previous list and says it is from the previous analysis. New searches use the changed membership and available analysis. Suggestions already shown can stay for up to ten days from first display, even after a large edit; their fit labels can still reflect the earlier playlist. Tracks you have added are filtered out once the edit is detected. Buffer suggestions are checked against the latest detected membership when requested. Candidate lists and Harmony can finish updating at different times.
If a suggestion is wrong for the playlist, mark it not a fit and it will not come back. Catalogue suggestions expire ten days after first display. Adding a track is up to you, in Spotify. Attracr notices when you do.
What is a candidate buffer?
Spotify does not expose your Liked Songs to any third-party app, Attracr included. So the tracks you save while listening — the ones you half-decided were good — are invisible to us, which is a shame, because they are the best candidates you have: you already picked them.
The workaround is a playlist you use as a staging area. When something catches your ear, drop it there instead of deciding on the spot. It is a pile of maybes, not a playlist you listen to.
Point Attracr at it under Settings, and suggestions from it carry a Buffer label in each tracked playlist's Candidates tab, showing which of those maybes actually fit that playlist. A drum & bass track you saved will surface under your drum & bass playlist and stay out of the acoustic one. You still add tracks in Spotify yourself; Attracr only tells you where each one belongs.
The buffer is meant to drain. Move a track into a playlist, or decide it was not as good as you thought, and delete it from the buffer — deleting it is how you tell Attracr to stop proposing it, everywhere at once. Nothing expires on its own, so a good track keeps showing up until you deal with it.
How does Attracr suggest tracks to remove?
A playlist that only grows slowly loses its shape. The same Candidates tab has a Consider removing list: a handful of tracks already in the playlist that are worth a second look. Each one is picked on two things, and the list says both in plain words:
- How long it has been in the playlist. A track gets a year before this counts at all. After that it counts a little, and from the third year on it counts more with every year. Keeping a playlist fresh means old tracks eventually make room.
- How close it sits to being an outlier. These are tracks that still belong, but sit near the edge of the playlist's sound: borderline on one feature, or clearly off on a feature that counts for little. Tracks that are already outliers are not on this list; the Harmony tab shows them.
So a track that has been there for years and sits near the edge is at the top, and one you added last month and fits well is not on the list. The list does not wait for a problem: it always shows the best few to look at, because how often you clean up is your call. When you agree, remove it in Spotify. Attracr cannot, and notices when you do.
What is a core track?
Core tracks are there to stay. They are the core of your playlist: they define it, its character, its identity. The song that set the mood when you started it, the one people follow it for. Its age says nothing about whether it should go. Mark a track as core, from the Tracks tab or on the removal list with the Off / 1y / ∞ switch, and Attracr will never suggest removing it. You can switch it back to Off from the Tracks tab.
Not sure a track should stay for good? Set the switch to 1y instead of ∞. After a year it drops out of the core by itself and is treated like any other track again.
Being core only affects that list. A core track that sounds different from the rest still shows under outliers on the Harmony tab, with the core icon next to it, because hiding it would make the harmony view dishonest.
Why do only some playlists show in the buffer picker?
Every playlist Attracr knows about has one of two roles. A tracked playlist is one you are analysing: it gets follower history, a harmony score, outlier detection and candidate suggestions. A buffer is a source of candidates instead — it is scraped daily so its tracks reach the pool, but it is never analysed on its own, and it does not appear on your dashboard. Analysing a pile of maybes would tell you nothing.
The picker offers only playlists you are not already tracking, because switching a tracked playlist to a buffer would stop its analytics, drop its harmony score and discard its candidate queue. That is too much to lose to a mis-click in a dropdown. If you really do want to reuse one, untrack it first and it will appear in the picker.
Why does my playlist have to be public?
Because that is the only version of it Attracr can see. A private playlist is invisible to a logged-out visitor, and Attracr is a logged-out visitor. You can make a playlist public, let Attracr read it, and set it back to private — though a playlist you are building for other people to find will need to be public eventually anyway.
Why not just use the official Spotify API?
Full Web API access is granted to applications with at least 250,000 monthly active users. That is not a threshold a new independent tool can reach before it has the access, so Attracr was designed to work without it. This is a constraint, not a complaint; it is also why Attracr's audio analysis is its own rather than borrowed.
Where do the audio numbers come from, if not from Spotify?
Spotify withdrew its audio-features endpoint in November 2024. Attracr analyses the recordings themselves and derives its own measurements — energy, valence, tempo and the rest. They are Attracr's numbers, which means they do not disappear when someone else changes an API.
What are the limits right now?
Attracr is in open beta and free. You can track up to five playlists. Limits will change as the analysis pipeline gets cheaper to run.
Is Attracr affiliated with Spotify?
No. Attracr is an independent product and is not affiliated with, endorsed by, or sponsored by Spotify AB or Google LLC.
Who runs Attracr, and how do I get in touch?
One independent developer, working from Belgium. Questions, bugs and data requests all reach the same inbox: hello@attracr.com.