Platform
The platform underneath every Spark ERP application.
Business applications, AI agents and the data platform all run on the same platform services — identity, workflow and integration — over shared, multi-tenant infrastructure.
Platform Overview
Spark ERP is built JSON-first: applications, pages, forms, workflows, rules and integrations are all persisted metadata artifacts, interpreted at runtime by a set of platform engines rather than compiled per tenant. That single design choice is what lets one platform run finance, HR, supply chain, manufacturing and service management as configuration, not forked codebases.
Metadata-driven
Pages, forms, workflows and rules are JSON artifacts, not per-tenant code forks.
Engine-based runtime
A fixed set of platform engines interprets that metadata identically for every tenant and app.
Multi-tenant from day one
Tenant isolation is a platform property, not something each application has to implement.
Architecture
Business applications, AI platform agents and the data platform all sit on the same layer, backed by shared platform services — identity, workflow and integration — over one infrastructure layer. See the full diagram and a layer-by-layer breakdown.
Multi-Tenancy
Every tenant gets its own isolated database, with per-application schemas underneath (ERP core, HCM, CRM, Finance and more). Isolation is enforced at the connection layer, not left to application-level filtering.
Security
Identity, single sign-on, MFA and role-based access control are platform services every application inherits, not something each team rebuilds. Tenant isolation, encryption and audit are enforced centrally.
Integration
A connector framework and API Orchestration Engine handle inbound and outbound integration — banking, payments, email, identity providers, payroll, e-commerce and more — as configured connectors rather than bespoke point-to-point code.
API Platform
Every entity, page, form and workflow exposed through Spark ERP is reachable through a versioned REST API secured by the same session and RBAC model as the UI — nothing in the product has UI-only access.
Scalability
The core deployable is stateless — any instance can serve any request, so the platform scales horizontally behind a load balancer rather than relying on sticky sessions or in-memory correctness state. Read-heavy workloads can be offloaded to read replicas independently of the write path.
Workflow
Approvals, multi-step processes and SLA-timed tasks run on a dedicated workflow engine — stateful process instances with human approval steps, not ad-hoc status fields scattered across modules.
Automation
Action chains let a form or page event trigger a sequence of steps — call an API, submit a form, navigate, notify — configured at authoring time. Scheduled jobs run the same way: one configuration-driven engine, not a bespoke job class per business need.
Data
One entity model spans every application. Fields, relationships and custom fields are designed once through the Entity Designer and are immediately available to forms, workflows, rules, reports and the API — the same record, everywhere it is used.
Extensibility
Beyond configuration, the platform accepts real code: a PF4J-based plugin runtime admits third-party plugins through the same block and API registration path the platform's own modules use, so an extension can never bypass validation the core respects.
Analytics
Dashboards and reports read live off the same entity model that powers the applications, so there is no separate reporting copy of the data to keep in sync.
AI
AI is a platform seam, not a bolted-on feature: an assistant, agents and automation sit alongside the business engines and can read and act on the same entity model, workflows and permissions every user already has.
Developer Platform
A CLI, SDKs and API let developers build applications, plugins, integrations and workflows against Spark ERP without touching platform source.
Deployment
Spark ERP ships as a single deployable suite for simpler operations, or as split platform and slice services behind a gateway for teams that need independent scaling — the same engines and metadata model either way.
See how the platform fits together.
Walk through the architecture, or talk to us about what a deployment looks like for your organization.