Built for Enterprise Scale
13 Laravel 11 microservices behind a Kong 3.9 API gateway, running on K3s Kubernetes. The same manifests that run the demo scale to a full HA cluster in any cloud or on-premise environment. No vendor lock-in. No managed services required.
Infrastructure
The full neXmis stack runs as Kubernetes workloads. The diagram below reflects the actual production layout including NodePorts, Kong routing, and per-service databases.
Under the Hood
Core Framework
Infrastructure
Data & Messaging
AI Engines
Security
Payments & Integrations
Enterprise Security
Security is layered at every level — from network ingress to database row. No single point of failure, no shared secrets.
TOTP-based 2FA on all accounts. SAML 2.0 for enterprise SSO. JWT tokens with short expiry and refresh rotation. No password stored in plaintext.
Role-based access control (RBAC) with per-module permission sets. Row-level scoping ensures users can only access their tenant's data.
Immutable, cryptographically signed event log. Every data read, write, export, and login is recorded with user, IP, timestamp, and entity reference.
GDPR-by-design. Consent management, right-to-erasure workflows, and Article 30 ROPA records generated automatically by Conform module.
TLS 1.3 enforced on all endpoints. cert-manager handles Let's Encrypt certificate rotation automatically. HSTS headers on all responses.
Full tenant isolation at the database level. Each organisation's data lives in separate schema scopes. Cross-tenant queries are structurally impossible.
All outbound webhooks are HMAC-SHA256 signed. Receiving systems can verify authenticity without trusting the network path.
Enterprise customers can deploy on their own infrastructure with no data leaving the organisation's network boundary. Air-gap deployment supported.
Data Isolation
Every tenant operates in its own MySQL schema namespace. The platform service enforces tenant context on every query via a global Eloquent scope. There is no shared table that could leak data between tenants.
Inside a tenant, users are assigned roles (Super Admin, Admin, Manager, Standard) with per-module permission sets. Roles are additive — you can build custom permission profiles without code changes.
All AI requests — whether to Claude, GPT-4, or Ollama — are assembled with a tenant-specific context snapshot. The AI never sees data from another tenant's records.
Enterprise deployments can have Kubernetes resource quotas applied per tenant namespace, ensuring one high-traffic tenant cannot starve another. CPU, memory, and storage limits are independently configurable.
Book a technical architecture call with our engineering team. We’ll walk through the full stack, answer security questionnaires, and discuss your deployment requirements.