Andrew G. Stanton - July 3, 2026
For the past week I've been thinking about one small but surprisingly important problem.
Where should Continuum store images?
At first glance, this doesn't seem like a difficult question. Plenty of services already host images.
Nostr clients commonly use Blossom servers. There are public image hosts. Cloud storage providers are everywhere.
But Continuum has never really been about choosing a particular service.
It's about making services interchangeable.
Images Are Part of Publishing
When writing an article, adding an image should feel natural.
Upload, insert, then publish
The author shouldn't have to think about where the file is stored or how it gets there.
That's the software's responsibility.
The First Backend
For the first implementation, I chose Cloudflare R2.
Not because it's the only solution.
Not because it's the most decentralized.
Simply because it solved the problem I had today.
It provides:
- Instant uploads.
- Stable URLs under my own domain.
- Generous free hosting for my current needs.
- No additional infrastructure to maintain.
- Storage completely independent of Git deployments.
An uploaded image immediately becomes available at a URL like:
https://media.mycontinuum.xyz/2026/pro/nostr/<npub>/images/...
No Git commit, deployments, no waiting.
The Backend Is An Implementation Detail
The interesting part isn't Cloudflare.
The interesting part is that the editor doesn't know—or care—which storage provider is being used.
Today the backend happens to be R2.
Tomorrow it could just as easily be:
- Blossom
- Amazon S3
- A self-hosted object store
- Local storage
- Something that hasn't been invented yet
The publishing workflow stays exactly the same.
Upload a file.
Receive a URL.
Continue writing.
More Than Image Hosting
The first version of Continuum Media already includes a small media library.
Images can be:
- Uploaded
- Viewed
- Deleted
- Organized with metadata
- Synchronized with the underlying storage
Instead of embedding information directly into directory structures, image metadata lives in a simple index.json file.
That makes it easy to add tags and additional information over time without changing the storage layout itself.
A Foundation
This isn't intended to compete with existing media hosting services.
It's intended to provide a foundation.
Just as Continuum separates identity from applications, it should also separate media management from any single hosting provider.
The storage backend becomes another interchangeable service.
That's a much more interesting problem to solve than simply uploading images.
Because in the end, Continuum isn't really about Cloudflare.
Or Blossom.
Or any particular technology.
It's about building software where your content, your identity, and now your media remain under your control—even when the underlying services change.
The goal isn't to make Cloudflare R2 permanent.
The goal is to make the storage backend replaceable.
Own your identity, your content and your media
The provider is simply an implementation detail.