> ## Documentation Index
> Fetch the complete documentation index at: https://docs.lazycloud.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Plans and limits

> What your plan includes, what usage costs, and what happens when you reach a limit.

Your plan belongs to your account and covers every workspace you own. The
dashboard manages it under Settings, then Billing, and the
[pricing page](https://lazycloud.dev/pricing) lists current prices and limits.

## What you pay for

Compute bills per second while a container runs, including startup, model
loading, and warm idle time. Volumes bill per GiB-month, and API responses and
internet egress are metered. Direct downloads through signed object-storage
URLs are free.

Every charge comes out of one prepaid credit balance, funded by trial credit,
subscription credit, and credit you buy. Purchased credit doesn't expire. Trial
and subscription credit expire on their stated dates. Subscription fees bill
separately. The Usage page breaks spend down by app and workload.

Idle charges come mostly from warm containers and running pods, so short
[`keep_warm`](/concepts/scaling#warm-containers) windows and stopped pods keep
them down. Stored files bill until you delete them.

## Automatic reload

Automatic reload charges your saved card when the balance drops to a threshold
you choose. A declined payment, or one that needs bank authentication, pauses
it until the payment method is updated. Buying credit, reloading, and changing
plans all need a saved card. A card alone unlocks your plan's GPU access and
concurrency, but adds no credit.

## Compute pricing

The pricing page lists base rates for automatic placement on preemptible
capacity, metered per second. One CPU is one vCPU. CPU and memory each bill at
the greater of the reservation and measured use. GPUs bill per card, on top of
the workload's CPU and memory.

Placement choices multiply the base compute rate:

| Workload setting                   | CPU and memory | GPU  |
| ---------------------------------- | -------------- | ---- |
| `preemptible=False`                | 3×             | 1×   |
| Pinned region or availability zone | 1.5×           | 1.5× |
| Both                               | 4.5×           | 1.5× |

Pinning placement needs a plan that allows it. Storage and transfer rates never
change with placement. Volume storage includes artifacts and prorates over a
30-day month.

On a connected AWS account, AWS bills the machines and LazyCloud charges a
management fee as a percentage of base compute rates. The
[pricing catalog](https://lazycloud.dev/api/v1/pricing) has every published
rate and its effective date.

## Plans

Free has no fee and a one-time trial credit. Team adds monthly usage credit,
higher concurrency, custom domains, and three members across the account's
workspaces. Business adds unlimited members, more credit and concurrency,
longer log and artifact retention, and a connected AWS account. Every plan can
join its own machines and use every GPU model, given a saved card, credit, and
availability.

Without a saved card, an account has lower concurrency and only the trial GPU
models. A plan change can't go below limits the account already exceeds, and
the dashboard names what has to go first.

## When you reach a limit

A container past the concurrency limit fails to start with a message naming
the limit, and existing work keeps running. Workspace and member limits apply
when those are created.

At a zero balance, new work and transfers stop and running compute shuts down.
Usage records periodically, so the balance can dip below zero first, and added
credit covers that before anything else.

A monthly usage limit, set in Billing, stops work the same way when reached and
resets on the first of each month at midnight UTC. It counts all credit usage
except subscription fees. Setting it to \$0 blocks new work.

If funds run out, stored data stays for 30 days, free of storage charges, then
is deleted. Credit that restores a positive balance in that window keeps it.