Skip to main content
To host an application on Vertra Cloud, log into the dashboard, click New Project, upload the code (a .zip or a GitHub repository), check what the platform detected, and deploy. The platform installs the dependencies, runs the build command if there is one, and starts the container.
Prefer the terminal? The CLI does the same with vertra deploy, and the VS Code extension deploys the folder open in the editor.

​
Create an account

The account is created on first login at vertracloud.app/login, with no separate sign-up form. You can sign in with Discord, GitHub, Google, or a magic link sent by email. The Free plan is currently “suspended”. While it existed, it was for testing the platform before using it in production: one application (no databases) with 100 MB of RAM “free”, and no web publishing. See what each plan unlocks in Plans and limits.

​
What does the New Project flow look like?

The New Project button opens a one-step-at-a-time flow. The first screen asks what you want to create and where the code comes from:

Upload a .zip

Drag in the project’s .zip. Folders like node_modules and venv are removed before upload.

Import from GitHub

Paste the repository URL or choose from your account’s repositories, including private ones.

Restore a snapshot

Create a new project from a snapshot of another project of yours.

Create a database

Managed PostgreSQL, MongoDB, Redis, or MySQL, with mandatory TLS.
At the top of the flow, you choose which workspace the project is created in (your own account or a workspace you’re part of) and, optionally, a folder to organize it in the project list.

​
Create an application

1

Source code

Upload the .zip or provide the GitHub repository. The project name comes from the file or the repository and can be changed here, along with an optional description.While reading the code, the platform already suggests the runtime, memory, and build and start commands. If the .zip has a vertracloud.config, its values take priority over the suggestion.
2

Build

Check the main file (chosen from a list of the project’s files), the runtime and version, and the memory.Advanced settings holds the start command and the build command. The build runs before start, on every deploy; if it fails, the application doesn’t come up. Next.js, Nuxt, SvelteKit, Remix, React Router, Astro, Gatsby and Vite projects already arrive with both commands filled in. If a declared dependency is missing, the step shows a warning without blocking you from continuing.
3

Environment

Add the environment variables (up to 25). PORT, HOST, PATH, HOME, USER and SHELL are reserved. Details in Configuration.
4

Network

Decide whether the application will have web publishing. With it on, the application gets a *.vertraweb.app address and needs to listen on port 80. Without it, the application runs in the background, like a bot or a worker.
  • Web publishing requires the Pro plan or higher and at least 512 MB of memory.
  • On Pro the subdomain is assigned at random. Choosing the name requires the Scale plan or higher.
  • A custom domain requires the Intermediary plan or higher.
Publishing can be turned on or off later, from the application’s Config tab.
5

Review and deploy

Review the summary and confirm. If the platform rejects a value (subdomain in use, memory above the plan), the flow goes back to the step where it’s edited.

​
What happens during deploy

  1. The code is uploaded to the platform.
  2. Dependencies are installed (such as npm install or pip install -r requirements.txt), with a 15-minute limit.
  3. The build command runs, if there is one, with a 10-minute limit.
  4. The container starts with the chosen memory and the start command.
  5. With web publishing, the subdomain is assigned and the TLS certificate is issued.
The .zip can be up to 100 MB. Creating an application, uploading code, and deploying via GitHub all count toward the plan’s deploys-per-hour limit.
Don’t upload node_modules, venv, .venv, __pycache__, .next, vendor, .bundle, target, .gradle or .local. The dashboard removes these folders from the .zip before upload, but they make the file bigger and the upload slower. Build output folders (dist/, build/) are not removed, because the main file is sometimes inside them.

​
Create a database

Choose the engine and the memory, then review. Storage is calculated from the memory and shown right below the control. The minimum memory is 512 MB for Redis and 1,024 MB for PostgreSQL, MongoDB and MySQL. Once created, credentials and certificates are on the database’s page. See Databases.

​
After deploy

The application’s page has these tabs:
  • Overview: status and CPU, memory, storage and network metrics.
  • Logs: the application’s output in real time, with platform events (restart and temporary pause by Vertra Shield, when applicable).
  • Files: the application’s files, with an in-browser editor (Pro plan or higher).
  • Activity: the history of what was done on the application and by whom.
  • Snapshots: create, restore and download copies of the project (Scale plan or higher).
  • Config: runtime, memory, start and build commands, environment variables, web publishing, subdomain, and custom domain.
To update the code, upload a new .zip, deploy via GitHub, or use vertra deploy. A plain restart doesn’t reinstall dependencies or redo the build; for that, use Reinstall dependencies or Force build.

​
Next steps

Configuration

Every vertracloud.config key, the editable fields, and environment variables.

Deploy via GitHub

Automatic redeploy on every push.

Tutorials

Guides by language, framework, and bot type.

Common errors

Symptom, cause, and fix for the most frequent deploy errors.