if a site hasnt been accessed in a while the first load might still be a bit depending on where you are in the world. i want to start tackling that now before i think of putting out more edge nodes (im not doing serverless and each node is a vps i control so its not that cheap)

wisp sites should feel even snappier as i did some work to drop TTFB, lemme know how it feels!

@upcloud.com is hardcore dropping packets for my singapore node and things have degraded badly. im trying to see what i can do but i may take it down temporarily and find a replacement to serve asia, pardon our dust

will have anycast soonish as bunny's geodns doesnt properly failover when a node is down

temporary downtime on us-west and asia nodes for maintenance on the nodes. will be done within 5 or so minutes

unlimited subdomains, file limits bump up to 300MB per file 700MB total

You can now create accounts with your github for easier onboarding

on the spirit of integrating better with other platforms here

an ai agent is already using it before i fully announced it and put proper docs for it up which is amazing to me

👀 soon

how do i acquire bubonic plague

i’m thinking i’ll have the private sites allow temporary anonymous uploads bound to two days maximum mostly so it can be used to quickly share things temporarily with the cli.

the regex parsers for permission sets just seems too strict and against spec, i have patches i need to look over before sending up

the old typescript binaries being 100MB each meant i was violating my own file limits i set for the service 😭

without bundling the runtime mem usage was 39 MB which is crazy for a js runtime! still pretty bad compared to compiled servers