How to do curation and businesses on Nostr

Suppose you want to start a Nostr business.

You might be tempted to make a closed platform that reuses Nostr identities and grabs (some) content from the external Nostr network, only to imprison it inside your thing – and then you’re going to run an amazing AI-powered algorithm on that content and “surface” only the best stuff and people will flock to your app.

This will be specially good if you’re going after one of the many unexplored niches of Nostr in which reading immediately from people you know doesn’t work as you generally want to discover new things from the outer world, such as:

But that is not the correct approach and damages the freedom and interoperability of Nostr, posing a centralization threat to the protocol. Even if it “works” and your business is incredibly successful it will just enshrine you as the head of a platform that controls users and thus is prone to all the bad things that happen to all these platforms. Your company will start to display ads and shape the public discourse, you’ll need a big legal team, the FBI will talk to you, advertisers will play a big role and so on.

If you are interested in Nostr today that must be because you appreciate the fact that it is not owned by any companies, so it’s safe to assume you don’t want to be that company that owns it. So what should you do instead? Here’s an idea in two steps:

  1. Write a Nostr client tailored to the niche you want to cover

If it’s a music sharing thing, then the client will have a way to play the audio and so on; if it’s a restaurant sharing it will have maps with the locations of the restaurants or whatever, you get the idea. Hopefully there will be a NIP or a NUD specifying how to create and interact with events relating to this niche, or you will write or contribute with the creation of one, because without interoperability none of this matters much.

The client should work independently of any special backend requirements and ideally be open-source. It should have a way for users to configure to which relays they want to connect to see “global” content – i.e., they might want to connect to wss:// to see only the latest music releases accredited by that label or to wss:// to get music from independent producers from that community.

  1. Run a relay that does all the magic

This is where your value-adding capabilities come into play: if you have that magic sauce you should be able to apply it here. Your service, let’s call it wss://, will charge people or do some KYM (know your music) validation or use some very advanced AI sorcery to filter out the spam and the garbage and display the best content to your users who will request the global feed from it (["REQ", "_", {}]), and this will cause people to want to publish to your relay while others will want to read from it.

You set your relay as the default option in the client and let things happen. Your relay is like your “website” and people are free to connect to it or not. You don’t own the network, you’re just competing against other websites on a leveled playing field, so you’re not responsible for it. Users get seamless browsing across multiple websites, unified identities, a unified interface (that could be different in a different client) and social interaction capabilities that work in the same way for all, and they do not depend on you, therefore they’re more likely to trust you.

Does this centralize the network still? But this a simple and easy way to go about the matter and scales well in all aspects.

Besides allowing users to connect to specific relays for getting a feed of curated content, such clients should also do all kinds of “social” (i.e. following, commenting etc) activities (if they choose to do that) using the outbox model – i.e. if I find a musician I like under wss:// and I decide to follow them I should keep getting updates from them even if they get banned from that relay and start publishing on wss:// or wss:// or whatever relay that doesn’t even know what music is.

The hardcoded defaults and manual typing of relay URLs can be annoying. But I think it works well at the current stage of Nostr development. Soon, though, we can create events that recommend other relays or share relay lists specific to each kind of activity so users can get in-app suggestions of relays their friends are using to get their music from and so on. That kind of stuff can go a long way.

Links to this page

This article on Nostr