# Free tier & pricing Every primary account includes an always-free allowance: one web service, one Postgres database, and one Redis database, each at the free unit of 0.5 vCPU / 1 GiB. This page explains how the allowance works, what the paid plans add, and how a monthly bill is composed. ## Free, for life Your primary account — the first account created for you at signup — can always run, at no cost and with no card on file: - - - Each runs at the **free unit**: 0.5 vCPU and 1 GiB of memory. There is no trial clock and nothing to cancel. > **Note:** > > This isn't a trial — the allowance never expires. Run the free services for a month or a decade; the price stays $0. ## How the free allowance works The allowance is a **price deduction, not a reserved shape**. Each free credit is worth the free unit's CPU + memory price and attaches to one running service of its type. At billing time the credit deducts `min(price(free unit), price(service))` from that service's monthly compute line: - - You get one free credit per service type, and it applies to your largest running service of that type — so the free share always lands where it saves you the most. No line goes below $0, and the credit doesn't carry over between months. Paid plans keep the same allowance — upgrading never takes the free services away. On paid plans, the plan fee is also granted back as a **monthly compute credit** that comes off your compute bill after the allowance; it doesn't roll over either. ## Worked example: upsizing your free service 1. 2. 3. The Free plan's ceilings are exactly the allowance, so step 2 requires a paid plan. Upgrading is self-service: pick a plan from the console's billing page and check out. ## Compare plans Plans differ in [seats](/docs/collaboration/teams-and-roles), ceilings (projects, locations per service, deployment history), and features (custom domains, auto-deploy from Git). Enterprise has no published price — talk to sales. Every number below renders live from the current pricing config: | Plan | Free | Plus | Pro | Enterprise | | --- | --- | --- | --- | --- | | Base price | $0 | $20/mo | $60/mo | $null/mo | | Seats included | 1 | 3 | 10 | 25 | | Projects | 1 | 10 | Unlimited | Unlimited | | Locations per service | 1 | 2 | 5 | 10 | | Free web service | 0.5 vCPU / 1 GiB | 0.5 vCPU / 1 GiB | 0.5 vCPU / 1 GiB | 0.5 vCPU / 1 GiB | | Free Postgres | 0.5 vCPU / 1 GiB | 0.5 vCPU / 1 GiB | 0.5 vCPU / 1 GiB | 0.5 vCPU / 1 GiB | | Free Redis | 0.5 vCPU / 1 GiB | 0.5 vCPU / 1 GiB | 0.5 vCPU / 1 GiB | 0.5 vCPU / 1 GiB | | Custom domains | — | 3 | 25 | 100 | | Auto-deploy from Git | — | Yes | Yes | Yes | | Priority support | — | — | Yes | Yes | | Log / metrics retention | 1 d | 7 d | 30 d | 90 d | What each ceiling means, what happens when you hit one, and the ceilings the table leaves out — environments, service counts per type — are all on [usage & limits](/docs/account/usage-and-limits). ## Service resources & prices Every service picks explicit CPU and memory from a discrete ladder, with a predictable monthly price linear in CPU + memory — see [resources & scaling](/docs/services/resources-and-scaling) for how sizing works. | vCPU | Memory | Price | | --- | --- | --- | | 0.25 vCPU | 512 MiB | $5/mo | | 0.5 vCPU | 1 GiB | $10/mo | | 1 vCPU | 2 GiB | $20/mo | | 2 vCPU | 4 GiB | $40/mo | | 4 vCPU | 8 GiB | $80/mo | Pick any CPU + memory (memory up to 8 GiB per vCPU). Databases price their compute the same way. A database's disk is derived from its memory — you never pick a disk size — and the table shows the disk that comes with each size: | vCPU | Memory | Price | | --- | --- | --- | | 0.25 vCPU | 512 MiB | $5/mo | | 0.5 vCPU | 1 GiB | $10/mo | | 1 vCPU | 2 GiB | $20/mo | | 2 vCPU | 4 GiB | $40/mo | | 4 vCPU | 8 GiB | $80/mo | Pick any CPU + memory (memory up to 8 GiB per vCPU). ## What a monthly bill looks like On the Free plan there is no billing at all: no subscription, no compute rating, no invoice, and no card. Billing begins only when you start a subscription on a paid plan. On a paid plan your bill has two parts. The **subscription** — the plan plus any seats beyond the plan's included count — is charged up front each period; the cadences and their prices are in the comparison table above. **Compute** bills at the end of each month, in arrears: each running service's monthly price × the fraction of the month it ran, minus the free allowance, minus your plan's compute credit. Two details worth knowing on paid plans: the free credit covers the free unit's **CPU and memory**, and **attached disk** bills per GB on top — for a database, the size tables above show the disk each size includes, priced into its monthly figure. You manage your plan and payment details from [billing](/docs/account/billing). ## FAQ ### Is it really free forever? Yes. The allowance is part of the pricing model, not a promotion — it has no expiry date and no usage clock. The free services keep running at $0 for as long as your account exists. ### What happens when I upsize? You pay only the difference above the free unit: the credit keeps deducting `price(free unit)` from the service's line no matter how large the service gets. Downsize back to the free unit and the service is fully covered again. ### Does a second account get free services too? No. The allowance belongs to your primary account. Additional accounts are payment-gated: they start without hosting, and every change is rejected until a subscription starts. See [multiple accounts](/docs/account/multiple-accounts). ### Do I need a credit card? No — signing up and running the allowance never asks for one. You add a card when you subscribe to a paid plan. ### What are the limits? The comparison table above lists each plan's headline ceilings; the full per-plan limits live on the usage & limits page. Separately, abuse limits are enforced: services used for cryptomining, attacks, or other resource abuse are suspended under the acceptable use policy. Ready to claim the allowance? Deploy your first service in the [quickstart](/docs/get-started/quickstart).