Skip to content
Scalpels home

Scalpels is officially launching on October 31, 2026. Special early adopter pricing is available now.

← All documentation

Documentation

Scalpels

Install a product

Updated

Browsing the catalog, reviewing requirements and installing are free. Starting an installation is available to team Owners and Admins.

Review the manifest

Open Catalog and select View requirements for a product. The installation screen shows the product's manifest: PHP version, Cloud resources, optional or required environment keys, build steps, and post-deploy commands.

Choose the Cloud region and, when the repository already has a Cloud application, an unused environment name. On a first install, Laravel Cloud creates a default environment with the application and Scalpels adopts that environment rather than creating a duplicate.

Installing a new Experimental product is for supporters: the Supporter tier, or any partner listing. Everything else, including redeploys and upgrades of an Experimental product you already run, is free.

Start the install

Before selecting Deploy, connect GitHub and Laravel Cloud. Scalpels then:

  1. Forks the source into the connected GitHub account, or reuses the matching fork.
  2. Finds or creates the Cloud application for that repository.
  3. Prepares the environment and configures the manifest's PHP version and build command.
  4. Creates only the database, cache, managed queue, object storage, or scheduler requested by the manifest.
  5. Waits for asynchronous resources to become ready and attaches them to the environment.
  6. Writes the manifest's application-owned environment values.
  7. Deploys and runs each declared post-deploy command.

The approved manifest and supplied values are snapshotted when the deployment starts. Supplied values are encrypted while queued and removed from Scalpels after they are delivered to Cloud. Completed steps and returned resource identifiers are checkpointed as the install proceeds.

Resume a failed install

A failure records its reason and leaves completed steps and created resources in place. Open the deployment page to inspect them.

After fixing the reported problem, an Owner or Admin can select Resume. Resume skips completed steps and retries the remaining work.