Use cases
Websites, game servers, self-hosted apps, build runners, databases. Each one turns on a different question, so we wrote them up separately: what we supply, what you run, and what to check before you pay.
01 / Stack control
A shop that crawls at evening peak rarely needs a bigger shared plan. It needs a box where you pick the runtime, the cache, and the deploy process.
Deciding question
Has the panel started getting in your way?
02 / Player routes
Hosting a server for a friend group scattered across China? Check the Linux server build, test both routes from their home networks, and read the AUP before you pay for anything.
Deciding question
Will the game run well for the players you actually have?
03 / App ownership
Vaultwarden, Paperless-ngx, a media server, a few containers on the side. Decide who keeps them updated and which data has to survive before you pick a plan.
Deciding question
Will you still be patching this app in a year?
04 / Code trust
A persistent runner with warm caches beats a cold container for most team builds. Keep it to your own repositories and give every job a cleanup rule.
Deciding question
Will only trusted code run here?
05 / Database exposure
The Postgres that outgrew shared hosting rarely needs its own server. Put it beside the app on one VPS, listener on loopback, and most of the exposure questions disappear.
Deciding question
Can the database stay off the public Internet?
Shared product facts
Every plan is the same unmanaged Linux KVM VPS underneath. We keep the server and network running; everything installed on it is your call and your responsibility.
Mainland-China paths are named rather than hidden behind an “Asia” label: China Telecom: CTGNet (formerly China Telecom CN2 GIA, AS23764 / AS4809); China Unicom: CUP (China Unicom Premium, AS9929 / AS10099); China Mobile: CMIN2 (China Mobile International N2, AS58807). You can check them from either Looking Glass before you order.
Fit depends on what you run
Root access and the plan table never change. The question worth asking does, and it depends on what you plan to run.
Current plan boundaries
Put your measured numbers next to the current vCPU, RAM, NVMe, and monthly transfer limits. A plan name tells you far less than that comparison does.
Looking Glass results show how the path looked at the moment you ran the test, not how your app will perform for users day to day.
Before you compare plans
Read the page that matches what you run, then hold your numbers against the current plans. Not sure it fits? Ask us; we would rather tell you no before you pay.
