How does Stormkit's self-hosted deployment model work?

Updated

Stormkit's self-hosted model means you run the entire platform — deployments, previews, environments, and the surrounding services — on infrastructure you control, rather than on Stormkit's cloud. You install it on your own server (on-premises or in your own cloud account), pick your own domain and settings during an interactive install, and keep the infrastructure and data in your hands. This suits teams that want to avoid vendor lock-in or need to keep data private. If you'd rather not run servers yourself, Stormkit also offers a Cloud option.

What "self-hosted" covers here

Stormkit describes itself as a self-hostable platform for running full applications. According to the site, a self-hosted install includes:

  • A managed database — the platform provisions and manages the database your app runs against.
  • Authentication — built-in auth for your applications.
  • Transactional email — sending app-triggered email (signups, notifications, and similar).
  • Cron jobs — scheduled tasks.
  • Analytics — performance, error, and user analytics.
  • Deployments, previews, and multiple environments — the core deploy workflow, with preview builds and separate environments.

The headline framing is "Deploy, Scale, and Own Your Web Apps. No Vendor Lock-In." — the point being that the platform layer is yours to host, not a service you rent.

How the install works

The documented path is a single install script:

curl -sSL https://www.stormkit.io/install.sh | sh

The site describes this as an interactive install: you pick the domain and settings as you go. So the expected sequence is:

  1. Run the install command on the server you intend to use.
  2. Answer the interactive prompts — domain and configuration settings.
  3. The platform comes up on your infrastructure, ready to connect a repository.

What to verify after install: that the platform is reachable at the domain you chose, and that the managed database and other bundled services started correctly. The install script handles setup, but the domain and settings are decisions you make during the run, so confirm they match what you intended.

Connecting a repository and deploying

Once the platform is running, the workflow is repository-driven:

  1. Connect your repository — the site lists GitHub, GitLab, and BitBucket as supported sources.
  2. Push to deploy — changes pushed to your repository deploy automatically.
  3. Each deployment gets a unique URL, with automatic TLS.
  4. Instant rollbacks — you can revert a deployment without rebuilding.

The site also advertises zero downtime rolling updates, so deploying an update doesn't interrupt users, and custom TLS certificates so you can use your own certificates for custom domains.

Where it runs

Stormkit's "Bring your own Cloud" positioning means you can run deployments on your own infrastructure — on-premises or in the cloud. It's also described as designed to integrate with existing cloud infrastructure, so you can keep using the tools and services you already have rather than replacing your stack.

The privacy angle is explicit: "Your data remains…" (the excerpt is cut off, but the framing is privacy-first, with data staying with you).

Enterprise-grade features worth knowing

Beyond the deploy loop, the site lists:

Feature What it does
Audit logs Track changes made by your team with detailed logs
Automated E2E testing Run automated scripts after deployments
Custom TLS certificates Use your own certs for custom domains
Zero downtime rolling updates Deploy without interrupting users
Monitoring Track performance, errors, and user analytics in real time

Who this model fits

Self-hosting Stormkit makes sense when:

  • Data privacy matters — you keep the data on your own infrastructure.
  • Vendor lock-in is a concern — the platform runs on your servers, not a vendor's.
  • You already have cloud or on-prem infrastructure — Stormkit is built to integrate with what you use.
  • You want the full platform, not just deploys — database, auth, email, cron, and analytics come with it.

If you don't want to operate servers, the Cloud option exists as the alternative. The site positions self-hosting as the path from side-projects up to enterprise scale, so the same model is presented as viable at both ends.

A note on the agent/MCP angle

The site mentions that an agent can run the whole platform through MCP — provisioning a server, deploying, publishing, and reading logs when something breaks. That's an additional way to drive the self-hosted platform, not a separate hosting model; the infrastructure and data still stay with you.

What isn't specified here

The provided material doesn't include pricing details, so whether self-hosting is free or paid isn't stated — check the Pricing and Enterprise pages before assuming. It also doesn't detail minimum server requirements or supported operating systems, so confirm those against the docs before provisioning.