podtunes logopodtunes

About

podtunes is a browser-based music player. Its library is populated exclusively from feeds tagged medium=music in the Podcasting 2.0 namespace and indexed on PodcastIndex.org.

Audio files are never re-hosted or downloaded. Every play button streams the original enclosure URL directly from the artist's feed.

Artists are identified by their itunes:author value combined with the feed's owner email, so multiple releases from the same artist collapse into a single profile page.

Playback is anonymous — no account required. Accounts are on the roadmap.

Found a bug or missing a feature? Send us a note via the feedback form.

Support podtunes

podtunes is free, has no ads and no tracking. If you would like to support the project, you can send a Bitcoin Lightning donation to our Lightning address podtunes@getalby.com. Any amount helps to keep the crawler running and the catalogue growing.

Music vs. talk podcasts

PodcastIndex's medium=music bucket occasionally contains talk shows that discuss music rather than release it. podtunes runs a lightweight classifier over every feed — combining feed categories, average episode length and title keywords — to keep the catalogue focused on actual music. Talk podcasts are hidden from every listing, chart and search result and do not influence the artist ranking.

AI-generated music

Some releases on podtunes carry an "AI" badge. We add this label when a feed explicitly discloses AI involvement through Podcasting 2.0 tags such as <podcast:ai> or <podcast:aiTransparency>, or when the feed metadata mentions known generative tools (for example Suno, Udio or Boomy) in keywords, generator fields or descriptions.

Since late 2026 we also run an audio analysis through SH Labs on newer releases. The service returns a probability that a track was generated by AI, plus a prediction such as "Pure AI Generated" or "Human Made". We only badge a release as detected AI-generated when a clear majority of its tracks exceed a high confidence threshold. The detector result is shown separately from artist self-disclosures, so you can always see which signal the badge is based on.

Both approaches are best-effort. Not every artist declares AI usage in their RSS feed, and automated audio detection can be uncertain with mastered or processed material. That means some tracks created with AI assistance may not show a badge, while a badge only ever reflects the strongest available signal. We treat the label as a transparency indicator, not as a verdict on quality or artistic merit.

Lightning boosts

Feeds that publish a <podcast:value> block get a Boost button on their artist and release pages. Payments are sent directly wallet-to-artist as Lightning keysends with Podcasting 2.0 TLV metadata — podtunes never touches the money and holds no keys.

The current build supports WebLN wallets such as Alby. Support for Nostr Wallet Connect (remote wallets like Alby Hub, Mutiny, Coinos, Primal) is on the roadmap so users without a browser extension can boost too.

How we rank artists

podtunes quietly computes an internal relevance score for every artist so lists like the artist directory can surface interesting profiles first. The score is not displayed anywhere in the UI — it only influences ordering. It combines the following signals, recalculated daily:

  • Charts presence — position on the Podcasting 2.0 V4V Music Chart (higher rank = more points).
  • Editorial recommendations — appearance and popularity on PodcastIndex's music recommendations.
  • Recency — how recently the artist released something, with an exponential decay of roughly a two-year half-life.
  • Release cadence — number of distinct releases in the last twelve months.
  • Catalog depth — total releases and total tracks, log-scaled so deep back catalogs are rewarded without dominating everything.
  • Collaboration graph — how often the artist is credited on other artists' releases via <podcast:person> or value splits.
  • Podcasting 2.0 adoption — small bonuses for a <podcast:value> block, a <podcast:funding> link, and a <podcast:person> profile image.
  • OP3 listener signal — 30-day download counts from The Open Podcast Prefix Project for feeds that route their enclosures through OP3, log-scaled so a few very popular shows don't drown out the long tail.

User-driven signals — plays, boosts, follows — will feed the score once accounts land. The score is deliberately opaque: it is a soft nudge for discovery, not a public leaderboard.