Pricing · beta
Pay for what you configure.
Only while it runs.
Three billing dimensions: CPU, memory, storage. That’s the whole list — requests are never counted and bandwidth is free, forever. Pick a size per function; we bill it per second while instances are awake and stop the meter at zero.
| CPU | Memory | Per second (running) | If it never slept |
|---|---|---|---|
| 0.1 vCPU | 128 MB | $0.00000035 | ≈ $0.91 / month |
| 0.25 vCPU | 256 MB | $0.00000085 | ≈ $2.21 / month |
| 0.5 vCPU | 512 MB | $0.0000017 | ≈ $4.42 / month |
| 1 vCPU | 1 GB | $0.0000034 | ≈ $8.82 / month |
“If it never slept” assumes one instance running 24/7 — most functions sleep most of the time, so real bills come out far lower.
Compute
The CPU and memory you configure, billed per second — only while an instance is actually running. Scaled to zero means $0.
$0.000003 / vCPU·s + $0.0000004 / GB·s
Storage
Source archives and built images for your functions.
1 GB free, then $0.02 / GB·month
Bandwidth
Traffic is not a billing dimension on Liusha. Outbound and inbound are both free — no meter, no cap, no overage.
Free. Always.
Free every month
no credit cardEvery account includes a monthly allowance that resets on the 1st. A typical hobby function never leaves it.
- 200,000 vCPU-seconds of compute per month
- 400,000 GB-seconds of memory per month
- 1 GB of storage
- Unlimited bandwidth — egress is always free
Why the bill stays small
Scale to zero is the default
You are never billed for an idle function. Most platforms charge for provisioned capacity; we only meter the seconds your code is actually up.
No per-request tax
Requests are free, however many you serve. A small instance handling thousands of requests costs the same as one handling ten.
Egress is free — really
The big clouds make their margin on outbound traffic. We charge $0 for it, with no meter and no cap. Serve files, stream responses, ship webhooks — bandwidth never shows up on your bill.
How billing works
Each function declares its CPU and memory size (default 0.1 vCPU / 128 MB, up to 1 vCPU / 1 GB). When traffic arrives, instances start and the per-second meter runs; when traffic stops, they scale back to zero and so does the meter. Storage is metered on what you keep. Requests and bandwidth are never billed. Beta pricing — existing users keep their rates when we launch.