I found Coracle before I found a reason to care about it.

What caught me wasn’t the interface. It was a single feature: the ability to add proof of work to posts. I’d read about NIP-13 on nips.nostr.com and thought it was clever. I just wished more people used it, because I’d never seen it anywhere. Until Coracle.

Seeing it running, actually doing something, produced a specific kind of excitement. The kind you get when a good idea finally gets its moment.

So I bookmarked it on my phones Home Screen, there was no native app. just because I wanted to use it. And suddenly I had four Nostr clients on my phone: one that was best for notifications, one that felt faster, one with the most features, and Coracle, because it had proof of work baked.

All of them exceptional, all of them the result of serious craft and genuine care, and all of them, if I’m honest, solving roughly the same problem. I don’t say that as a criticism. I say it as a question worth sitting with.

Then the amazing Nostria arrived, and SondreB added a fifth. But this one stopped me cold for a different reason. Not a new feature to collect, but a realisation: this protocol is so deep you could build an entire internet on top of it. Exploring it felt like stumbling onto a world under active construction, the way the Earth is being secretly rebuilt at the end of the Hitchhiker’s Guide, piece by piece, with quiet intention and enormous ambition hiding behind ordinary surfaces. I don’t think there is much you cannot do in Sondre’s Nostria.

Then I stopped and asked myself: how did I get here, and why did it get me excited?

I got here because I read a protocol spec. Most people never do that. They find a product because it solves something for them, and if they’re curious, they eventually lift the hood. But very few people fall in love with the engine first and go looking for a car to put it in. Builders do. Users don’t.

And if we want Nostr to spread, really spread, we have to build for users. That means resisting the urge to make another showcase. Another client that demonstrates as many NIPs as possible, another app where “Nostr-native” is the main feature. These things are exciting to us as builders and protocol enthusiasts, but they’re not the entry point for anyone else.

The entry point is whatever problem someone actually has. A tool that helps people find things near them, a way to coordinate a community, a place to share work and get feedback. Build that well, and quietly wire Nostr underneath it. Users don’t need to know, they need the problem they are having resolved. And then, six months later, they discover that the account they made on your app works somewhere else entirely. That they can take their history, their identity, their reputation with them, and that no single company owns any of it.

That feels like magic, and that’s when Nostr clicks for a normal person. Not when you explain it to them up front. I’ve been experimenting with this on two small test projects that are nowhere near ready for the world. Partly because I’m not the best hustler and partly because I may be infringing on some of the core ideas behind nostr (more on that later). but I secretly think that’s ok. Nearhood.co.uk is an attempt to connect people with their immediate neighbourhood: hiring the plumber next door instead of PlumbCo, finding the pub quiz that the security guard from Tesco also goes to. The kind of local fabric that used to exist before the internet flattened everything into platforms. Agentlist.com is a different experiment: a place to share agent skills in a usable, discoverable way. There’s a listing on there that defines how to write prose that’s actually enjoyable to read, and I used it to help write this one. Emphasis on help, as I hate “generatia” as much as the next guy. I mean what I write here. I hope it works and you are with me so far. In both cases, nearhood.co.uk and agentlist.com Nostr sits underneath, handling auth and distribution (and for nearhood, messaging), and neither app makes a point of saying so. I also, and here is maybe where people might disagree with my approach, used passkeys to deterministically generate a key pair for identity. Secured by your enclave on your device, and you don’t need to remember anything, unless it’s been 6 months and you discovered you can take the key with you. Then write it down.

Then: timing. because the timing matters enormously. We are in a moment where large language models can generate enormous amounts of code at speed. The danger is that everyone will implement the same basic things differently, each version slightly broken and each one a new surface to hack. Auth reinvented badly. Message formats that don’t interoperate. Distribution bolted on as an afterthought. Nostr arrived at exactly the right moment to short-circuit all of that: a clear spec for how messages should be structured, a working battle-tested approach to auth, and a distribution model that’s resilient by design and hard to compromise. Solid infrastructure for the next decade of software, already here, already proven. The builders who see this clearly aren’t the ones making more demos. They’re the ones quietly wiring Nostr into products that people will use for completely ordinary reasons, discovering later that they’re holding something much more powerful than they expected. There are people out there who don’t know they need Nostr yet. Let’s bring it to them.​​​​​​​​​​​​​​​​

Written on Nostria,

Mostly by hand and with help from this agent