Dating for people who actually listen
Your feed says more about you than your bio ever could. The shows you binged in a weekend, the backlog you guard like a reading list, the episode you've sent to every person you've ever cared about. That is taste. That is identity.
Try the demoHow it works for podcast people
Connect via Spotify. We pull your podcast subscriptions, listening history, and episode completion data. Nothing is shared raw - only derived signals. All opt-in, all revocable.
The algorithm understands that podcasts exist in layers:
- An episode belongs to a show, made by a creator, tagged with categories
- Matching happens at every level - shared shows, shared categories, shared listening patterns
- If you listen on Spotify and they use Apple Podcasts, the same show resolves to the same entity
- If we have rich episode-level data but no explicit category affinity, we compute it from your subscriptions
Niche overlap > mainstream overlap
Both subscribing to The Joe Rogan Experience? That is a popularity 95 overlap. Half the internet listens to it. It barely registers. Both deep into The Magnus Archives? That is a signal.
But it goes further. The algorithm knows the difference between "subscribed but never finished episode one" and "listened to all 200 episodes, twice, and cried at the finale." Engagement depth and completion rates turn a flat subscription overlap into a real affinity score.
The core signal
A subscription list is not a personality. Anyone can hit follow. What matters is how you listen.
Did you complete the entire back catalogue in two weeks? Did you stop at episode three and never come back? Did you relisten to specific episodes? The algorithm captures engagement depth - minutes listened, completion rate, binge velocity - and uses it to separate genuine fans from passive subscribers.
Sentiment matters too. On platforms that surface it, saving episodes, adding shows to queues, and coming back to a show after months away are all positive signals. Subscribing and immediately dropping off is not.
What gets scored
Example match
Not "you both like podcasts." The actual data. The actual niche.
The backlog says everything
Sharing a love for a show with 500 subscribers contributes 17x more signal than sharing one in the Spotify Top 10. The algorithm is built for people whose podcast feed is a curated identity - not background noise.
That horror fiction podcast you recommend to everyone. The linguistics show you found through a Reddit thread at 1am. The 6-season investigative series you listened to twice. Those are the data points that find someone who actually gets it.
And it works both ways. Two people who both dropped the same massively popular show after three episodes? That shared avoidance of mainstream content is weighted by niche rarity too. Bonding over what you don't listen to is just as valid.
Integrations
All opt-in. All revocable. We show signals, never raw data.
Affinity Atlas is in development
No real matching is live yet. If you want to be the first to know when it launches - or you have strong opinions about how podcast taste should factor into dating - get in touch.
Stay in the loop