Edge nodes: one today, and the knob for when there's more

Short, honest post: the hosting plane is still one node. No secret map of PoPs lit up behind a curtain — the edge story is aspirational until a second node joins the first. What is pinned is how you'll scope a site when there's more than one place to scope it from, and that's worth writing down before the scaling makes it rushed.

The scoping knob

The default, once there are multiple nodes, is "replicate everywhere". The edge pragma opts a site into a subset:

routes:
  - path: /*
    edge: region:us,eu   # only US and EU nodes

Useful when replication to a region serves traffic you don't have, or when data-residency constraints pin a site to a specific set of jurisdictions.

How regions get picked

No roadmap here — calling the rollout order in advance would be a strategy claim, and the honest answer is: we'll follow the users. Deploy a site, tell us where your visitors are, that's the input. The initial additions go wherever sites are asking from, and growth from there is gradual rather than a single push.

What else is coming

• Edge compute — small functions at the node level (auth checks, redirects, header transforms). The edge: route field is where they attach.
• Per-site traffic views — the stats sparkline on the demo dashboard is the first tile; proper per-region breakdowns come after multi-node does.

The shape stays small: a config file, a command, a deploy. Enough coverage to make indie sites feel close to their users, not enough to need a sales team.