Vertra Cloud Plans and Limits

Vertra Cloud is a Brazilian application hosting platform (PaaS): deploy from GitHub or a zip file, isolated containers, managed databases, file storage, snapshots, a CLI, and a VS Code extension. This page gathers, in one place, every public number of the platform — price, memory, vCPU, projects, and API limit per plan. The values below are the same ones used by the platform in production. Prices are in Brazilian reais. You can buy 1, 3, 6 or 12 months at a time: 3 months get 10% off, 6 months get 15% off and 12 months get 25% off. Pay with PIX or credit card (Visa and Mastercard) on any plan. A plan paid by card renews automatically on your default card, 3 days before it expires; use Cancel renewal in Billing to stop it. PIX payments never renew by themselves. The prices below are the monthly price.

​

How much does each Vertra Cloud plan cost?

Vertra Cloud has 22 plans in four families: Starter (1 and 2 GB of RAM), Pro (4, 6 and 8 GB), Business (12, 16 and 24 GB) and Enterprise (14 sizes, from 32 GB to 512 GB). Plans range from R$7.70/month (Starter 1) to R$2,911.90/month (Enterprise 512, 512 GB of RAM). Memory is the only thing apps and databases consume — there’s no per-hour or per-request charge. There is no project count cap on any plan — the real limit is the plan’s RAM: a creation is refused with MEMORY_LIMIT_EXCEEDED once the plan’s RAM runs out, so a 100 MB app fits 10 times on Starter 1. All plans include Vertra Shield, the platform’s traffic protection. When an application is paused, the dashboard shows the reason and the remaining time. The bandwidth limits by memory, automatic TLS certificates, and 99.99% SLA also apply to all plans. The “API requests/day” column above only applies to calls authenticated with an API key (dashboard usage doesn’t count) and resets at midnight in the Brasília time zone. The same limit also has a per-minute window. An account with no active plan has 30 requests/min and 1,000 requests/day. Details on both windows, response headers, and error codes are in API Limits.

​

How much vCPU does an application actually get?

The vCPU column above is the ceiling per application on the plan, not a total split across the account’s projects. Within that ceiling, a project’s vCPU grows with the project’s RAM:

• Growth: on plans with up to 2 vCPU, +1 vCPU for every 512 MB of RAM; on plans with more than 2 vCPU, +1 vCPU for every 1 GB (1,024 MB).
• Floor: a project with 512 MB or more gets at least 2 vCPU; below 512 MB, at least 1 vCPU.
• Ceiling: never more than the plan’s vCPU. Starter 1 (1 vCPU) gives every project 1 vCPU.

The same rule applies to databases. More RAM never results in less vCPU. A change takes effect the next time the project restarts or is deployed: when your plan changes, the new ceiling reaches each application and database on its next restart or deploy, and nothing restarts by itself.

​

What happens to an account without an active plan?

An account without an active plan can’t create applications or databases or publish a site: those requests are refused with PLAN_REQUIRED (403). Subscribe to a plan, starting at Starter 1, to continue. Projects that stay offline for more than 30 days on an account without an active plan are automatically deleted, with a snapshot taken first and email warnings on days 7, 23, and 27.

​

Which features does each plan unlock?

Every feature has a minimum plan, and it applies to the resource’s owner — inviting someone on a higher plan into a workspace doesn’t unlock the feature. Publishing an application to the web requires at least 512 MB of RAM allocated to it. The “Powered by Vertra Cloud” badge appears on the pages of sites on Starter 1 only.

​

How many custom domains and ports does each plan allow?

Custom domains start at the Starter 2 plan. The limit is per application and applies to the application owner’s plan. The automatic www (or root) companion that comes with a domain does not count. Beyond the plan’s domains, every application with domains can serve them on up to 5 different ports besides port 80; each domain also accepts an optional path prefix and an optional redirect to another domain of the same application. Reaching the domain limit answers DOMAIN_LIMIT_REACHED, and the port limit DOMAIN_PORT_LIMIT_REACHED. A domain that stays pending, moved or deleted for 14 days is removed automatically; after any removal, delete its CNAME and TXT records at your DNS provider. See Connect custom domains. Templates follow the same limits: each one needs a minimum plan, calculated from what it creates (applications, databases, Blob, image or web publishing), and image templates use image slots, which start at the Pro plan. The deploy tells you what is missing before it starts.

​

How do I connect AI tools to Vertra Cloud?

Connect your compatible agent through the MCP server to operate your account’s resources.

​

Which languages and versions does Vertra Cloud support?

The platform builds and runs applications on 13 runtimes. Each one has a recommended version, a latest version (often the same one) and older versions you can pin by number — in the wizard, in Settings → General, in VERSION or through the API. Older versions are removed when they reach end of support; an application saved on a removed version runs on the newest release of the same line, or on recommended if the whole line was removed.

​

Which databases are managed?

There are four managed engines, all with mandatory TLS and certificates (CA, server, and client) generated automatically by the platform. You choose the version when you create a database (the dashboard selector, or version in the API); the recommended one is the default. Existing databases keep their major version, and security updates (patch releases) are applied automatically whenever the database is restarted (stop, then start, in the dashboard). If you use btree_gist indexes on float columns or ltree indexes in PostgreSQL, run REINDEX on them after the restart. Existing MySQL 8.0 databases stay on 8.0, which isn’t available for new ones. Existing Redis 7.4 databases stay on 7.4. A database created from a snapshot is born on the snapshot’s major version (a MySQL 8.0 snapshot creates a MySQL 8.4 database, and a Redis 7.4 snapshot a Redis 8.2 one); version can be the same or newer, never older (SNAPSHOT_VERSION_NEWER). Restoring a snapshot taken on a newer version into an older database is refused with the same code and nothing is changed. vertracloud is the user of databases created from v26.4.8 on, and it isn’t a superuser. Older databases keep the engine’s user: postgres (database postgres), default on MongoDB (auth database admin) and Redis, and root (database default) on MySQL. The username field of the database in the API tells which one to use. The port is unique per database — not the engine’s default — and is shown on the database page. See how to connect.

​

What are the limits of the Data tab?

The Data tab browses and edits database content through the dashboard or the API, without exposing the password or the connection string, and requires the Pro plan or higher. Below that plan, the API responds with PLAN_RESTRICTED_FEATURE (403). Reading applies to any workspace member; writing and the SQL console are owner or admin only. Administrative commands and arbitrary code execution are refused in the Redis and MongoDB console.

​

How much Blob Storage does each plan include?

Blob Storage is available on every plan. The quota counts every stored file of the workspace owner. Folder rules, custom domains and image transformations also depend on the plan. The full rules are in the Blob Storage section of the knowledge base. Above the quota, uploads are blocked. Usage above the quota can only come from a downgrade, and it is charged at R$0.20 per started GB on renewal. When a plan expires, there are 72 hours of grace before the files are deleted.

​

How many members, Docker images and manual backups does each plan allow?

Workspaces start at Starter 2, custom Docker images at Pro. Manual backups have two daily limits, counted on the calendar day in Brasília time: up to 5 per project (application or database) and, across all your projects, 2 per GB of plan RAM per account (at least 5, at most 64), shown in the table; the automatic and safety backups don’t use them. A pending workspace invite counts as a member.

​

What are the deploy and resource limits?

Network bandwidth is proportional to allocated RAM: 50 Mbit at 100 MB, 100 Mbit at 256 MB, 200 Mbit at 512 MB, 400 Mbit at 1,024 MB, 800 Mbit at 2,048 MB, and 1,600 Mbit at 4,096 MB.

​

Deploys-per-hour limit

Counts as a deploy: creating an application, uploading a zip, deploying via GitHub (manual or by push), reinstalling dependencies, and changing the build command. Restarting and stopping don’t count. A GitHub push that arrives past the limit is ignored and logged in the activity history.

​

Deploy history and rollback limits

Each application keeps the code of its previous versions for a limited time, so you can roll back to one of them. The version that is running now never expires. A previous version’s validity counts from the moment it stops being the running one.

​

Automatic rollback and zero-downtime deploys

Automatic rollback (Pro and above): if the new version of an application stops within 60 seconds after a deploy, the platform goes back to the previous version once and notifies you. It is off by default; turn it on in the application’s settings. Zero-downtime deploys (Starter 2 and above): the new version starts next to the running one and takes over once healthy. It is off by default (off); auto applies it to applications that receive web traffic and on to every application. See Update Application Configuration for the setting.

​

Web Analytics history

Web Analytics keeps a queryable history per application. The summary, traffic and errors views are on every plan; the performance and requests views start at Pro.

​

Next steps