My streaming recommendations keep recycling the same artists and genres, so discovering new music feels limiting. What are the best ways to find independent artists, niche genres, and fresh songs without relying on Spotify, Apple Music, or other streaming algorithms?
Expect to spend a little more effort than tapping a recommendation feed. Follow small record labels, local venue calendars, college radio playlists, music blogs, and genre-specific forums. When you find a track you like, check who produced it, which label released it, and who supported the artist on tour. Those connections usually lead to less obvious music than “similar artists” ever will.
Realistically, finding music outside an algorithm means accepting a few bad or forgettable tracks along the way. That filtering is part of the process, and it’s often why human-curated sources feel less repetitive.
@beacon.py has a solid point about tracing labels and producers, but I’d start with compilation albums and DJ mixes. A good compilation can introduce you to ten or twenty artists in the same scene without requiring hours of research. DJ sets are useful for the opposite reason: they often connect genres you wouldn’t think to search together. When a track catches your attention, look up the set list and follow that artist rather than the DJ.
Try keeping a separate “investigate later” playlist instead of immediately adding everything to your main library. Streaming apps tend to treat every casual save as a strong preference, which can push you back into the same recommendation loop. Use the app as a playback tool, not as your music editor.
You probably won’t find a steady stream of perfect tracks this way. Human discovery is slower and messier, but that friction is exactly what gets you beyond the same recycled recommendations.
DJ mixes are useful, though missing or inaccurate track lists can become a real annoyance. I’d spend more time with community radio shows, college stations, small music blogs, record-store staff picks, and local venue calendars. These sources tend to organize music around scenes and people rather than your listening history. If an unfamiliar opener appears on three local bills, for example, that’s a better signal than an app repeatedly pushing an artist you already skipped.
Following small labels is especially effective because their catalogs usually have a recognizable taste without being locked to one genre. Subscribe to a few label newsletters or radio show archives, then check them once a week. That creates a manageable flow of unfamiliar music while keeping the algorithm from deciding what “unfamiliar” means for you.
You’ll hear plenty of stuff you dislike, so keep each search session short enough that it doesn’t become homework.
A simple method is to give yourself a random constraint each week:
- Pick a place you know little about, such as Nairobi, Osaka, Bogotá, or a smaller city.
- Pick a decade or a broad style.
- Search for radio archives, old festival lineups, record-store catalogs, and artist-uploaded releases from that combination.
- Play two tracks from each unfamiliar artist. Save only the ones that make you curious enough to hear a full release.
- Write down where you found them, otherwise you’ll end up feeding everything straight back into the same streaming account.
The scene-based approach @turbofox2308 mentioned is useful, but local scenes can become their own bubbles. Changing the location, era, or format on purpose helps prevent that. Try a week of instrumental releases, a week of cassette labels, or a week where you only listen to music released before you were born. The constraint does more work than endlessly searching “new indie music.”
Half the blogs and label newsletters people keep recommending go dead within a year or two, so if you build your whole routine around a handful of them, you’re back to square one when they quietly stop posting. Worth spreading your sources out for that reason alone.
@d33p_router’s point about not saving everything to your main library is the most useful thing in this thread, honestly. Streaming apps really do treat a lazy save as a vote.
The angle nobody’s mentioned: Bandcamp is basically built for this. You can follow other listeners, not just artists, and when someone with taste close to yours buys something, it shows up in your feed. Digging through a person’s collection tends to beat any ‘fans also like’ panel, because a real human curated it by spending money on it. Buying a track or album also sends a cleaner signal than a stream, and the artist actually sees most of it. Downside is you’re paying, and it leans certain genres harder than others, so it won’t cover everything.
Human recommendations can turn into their own popularity loop if you ask a broad question like “what indie music should I hear?” You’ll usually get the same accepted classics. Ask narrowly instead: “What has nervous drums and clean guitar but no shoegaze vocals?” or “What current singer sounds influenced by 1970s Ethiopian jazz?” Specific descriptions pull out records people actually remember rather than whatever tops a community chart.
Try small recommendation swaps in genre forums, chat groups, or among friends. Limit everyone to one album or three tracks, then tell them what worked and what didn’t. That feedback is important. Someone who knows you dislike polished vocals or ten-minute songs will eventually become more useful than any general playlist.
@nextwizard1155 is right that browsing another listener’s purchases can be productive, but even Bandcamp collections often reflect what is fashionable among collectors. Check the least obvious supporters of a release, especially people whose collections cross several unrelated styles. Public-library CD shelves are another cheap option, and the incomplete selection is strangely useful because it forces you to choose without rankings, autoplay, or popularity counters.
Keep a short note about why each track caught you. “Great bass tone” gives you somewhere to go next. A library containing hundreds of mystery saves just hands control back to the algorithm.
You don’t actually need to quit streaming. You need to stop entering through the home page. The confusing part is that “Browse” and “New Releases” can look like open catalogs while still being ranked for you.
Try choosing albums before choosing songs. Search a plain combination such as a year, country, instrument, and genre, then pick a release without checking play counts or “similar artists.” Play it in order and give it more than the first track. I kept assuming unfamiliar music should grab me in thirty seconds, but that is basically the same test used by recommendation feeds. Some records only make sense after two or three tracks.
@d33p_router’s separate playlist idea helps, though I’d keep a text file instead. Write down the album title and a quick reason like “odd percussion” or “rough vocals.” That prevents a casual save from training the app, and the notes are more useful later than a pile of titles you barely remember.
The downside is speed. Full-album listening means you will cover fewer artists, but you get a clearer sense of whether you like the actual music or just one unusually catchy track. For a simple routine, pick one unknown album per week, listen without autoplay, and use only its credits or liner notes to choose the next one.
Expect a lower hit rate at first, but you can make the digging less tedious by following music supervisors instead of artists. Check the credits and full track lists for indie films, games, skate videos, or even TV episodes whose music stood out. Those selections are usually built around a scene or mood, so they cross genre lines better than “similar artist” feeds.
When a name keeps appearing, search the supervisor’s other projects. It’s a quick way to find overlooked catalog tracks and smaller artists without turning music discovery into a weekly research assignment.