Elementary Interactive

The layer underneath

Every property your organisation runs, on one platform, with one team responsible for keeping it current.

At its core sits the Central Admin: one interface to manage users, permissions, content and integrations across all your sites and applications. No scattered logins, no version mismatches, no “which plugin handles that?” moments.

Architecture

Built as separate packages, not as a monolith split into folders.

Modular by construction

Each module declares its own dependencies and can be installed, updated or replaced independently. Customisations live in their own packages, so they survive upgrades instead of being overwritten by them.

Thin front ends

Websites contain templates and their own versioned packages. Business logic stays in the platform. A design change doesn't touch the system underneath, and a platform update doesn't require you to redeploy.

Shared foundation

One running unit serves a site; the platform base is shared across all of them. The marginal cost of the next property is small, which is why an estate of twelve isn't twelve systems.

Configuration, not code

Data structures, fields, taxonomies, permissions and workflows are records in a database. A new requirement is usually a setting, not a release.
Technical specifications

Isolation and control

Every request is scoped to a company and a site, enforced below the application.

Database-level tenancy

Isolation is enforced at the database level, not only in application logic. Each tenant can run on a dedicated database, with credentials encrypted at rest and decrypted only at connection time.

Consistent across everything

Tenant scope applies identically to HTTP requests, background jobs, CLI commands and WebSocket connections. A queued job cannot reach another tenant's data.

Granular permissions

Roles assembled from named rule groups, down to individual content types or specific fields. A marketing team member can manage one site's news without seeing another's.

Attributable usage

Users belong to companies, so usage is separable by organisational unit and can be recharged internally.
GDPR compliance

Staying current

Security fixes appear in the actively maintained major versions of the components a platform is built on. A system that has drifted onto an unmaintained version has two options when an advisory lands: backport it, or migrate under pressure. Both happen at the worst possible moment.

Nitrogen runs on actively maintained foundations, which makes a security fix a routine package update rather than a project. A significant share of what would otherwise be third-party packages is replaced by our own — a smaller dependency surface, and a fix time that is ours to control rather than someone else's to schedule.

The trade-off, stated: this makes the system less open and less reliant on the development capacity of the open-source community. In production, predictable remediation matters more. We would rather say that here than be asked about it later.

How we work

Operations

The platform is operated, not handed over for you to run.

Zero-downtime deployments

Updates roll out without service interruption. Your visitors never see a maintenance page.

Staging environments

Test changes safely before they touch production.

Monitoring, not reporting

Availability and performance watched continuously, and infrastructure failures handled when they occur rather than when they are reported.

Backups

Automated, tested and restorable to a point in time, with encrypted off-site copies.
Hosting

Audit

Every model change is recorded automatically: who changed it, when, from which IP address, and what exactly changed, field by field, with the previous and new values. Logins and logouts are recorded with IP address and user agent.

This is not a log you enable. It is how the platform stores changes, which is why the answer to “who changed this, and when” takes seconds rather than a forensic exercise.

What transfers to you, and what doesn't.

What you own