Platform  /  Architecture

One data spine. Zero silos.

The contract is the central record connecting every department. When support opens a ticket, they see the billing balance and signal levels instantly. No integrations, just one unified database.

300+ DB tables 120+ API endpoints Multi-tenant kernel Self-host or SaaS Configuration-driven v3.7.0 · Live
The principle

What "one data spine" actually means.

In most ISP operations, the customer exists in five places: as an account in the billing system, as a device record in the network management tool, as a ticket in the helpdesk, as a line item in the inventory system, and as a row in a finance spreadsheet. Each system has its own version of the truth. When these versions disagree, and they always do, someone has to reconcile them manually.

ISPCQ eliminates this by design. There is one customer record. One contract. That contract connects to the billing ledger, the installed equipment serials, the OLT port assignment, the signal history, the support ticket timeline, and the payment history.

This is not an integration layer sitting on top of separate systems. It's a single application with a single data model. The billing engine writes to the same ledger that finance reads. The network provisioning system reads the same contract that support sees.

There is no synchronization delay, no mapping table, and no "sync failed" error at 3 AM.

Architecture

Built for scale and sovereignty.

A battle-tested stack designed for operational reliability. Multi-region, multi-tenant, AI-native, configuration-driven.

01
Multi-region ready
Address / currency / tax / i18n by configuration

One codebase, infinite localizations. An ISP group running operations in multiple countries doesn't need separate installations or code forks. Regional differences (address formats, tax rules, currency display, payment gateway configuration) are handled through data-driven configuration layers.

  • Address engine: database-driven templates ("Street + No" vs "Complex + Unit")
  • Currency: configurable formatting and display rules
  • Tax: pluggable VAT/tax modules per region
  • Translation: database-driven i18n; all labels and notifications via __T() lookups; add languages without code changes
02
Operational foundation
Stable runtime + cron engine + multi-tenant kernel

ISPCQ runs on a proven stack optimized for ISP transaction volumes: thousands of billing cycles, hundreds of concurrent users, and continuous network polling, all on modest server hardware. Designed for operators who want fast page loads and low maintenance, not engineering teams managing microservices.

  • Stable runtime built for sustained high-volume transactional work
  • Unified data layer; billing, network, inventory share one operational truth
  • Cron engine: billing cycles, DDO processing, signal polling, ONU monitoring, report generation on configurable schedules
  • Multi-tenant kernel: one codebase serves multiple tenants; differences are configuration, not code forks
03
Aelita AI integration
Embedded LLM with permission-scoped access

Embedded LLM intelligence powered by Claude. Aelita operates within the same permission boundaries as the logged-in user. She can draft, suggest, and analyze, but she cannot act without human approval. AI acceleration without AI risk.

  • Ticket copilot: drafts responses based on contract / device / billing context
  • Billing analyst: explains invoices and detects anomalies
  • Docs vision: extracts data from supplier invoices and contracts
  • Twenty-one capabilities shipped today; see the full list
04
Hot-pluggable extension architecture
Modules & plugins switched on per-tenant by configuration

Payment gateways, OLT and network drivers, and messaging channels ship as self-contained, manifest-described extensions, switched on per-tenant through database configuration. No code forks, no core edits. Each extension declares its own boot files, dependencies, and config keys, so an operator's capability set is composed from configuration rather than custom builds.

  • Two-tier loader: full modules load first, small plugins second, so dependencies resolve cleanly at boot
  • Database-driven activation: the active set is read from per-tenant config; the same codebase exposes a different capability mix per operator
  • Manifest contracts: every extension names its boot files, required dependencies, and config section/key in one declaration
  • Composable catalogue: payment gateways, OLT/network drivers, and messaging channels added or removed without touching the core
Real-life scenario

The contract as the nucleus of every decision.

Scenario. Customer #4721 calls to cancel their service. In a disconnected stack, cancellation is a multi-step process across four systems: the support agent creates a cancellation request, someone in billing has to stop the next invoice, the NOC team needs to deprovision the OLT port, and the inventory team needs to schedule equipment recovery. If any of these steps is missed, the customer gets billed after cancellation, or the ONU sits at their premises for months, or the OLT port stays allocated and unavailable.

In ISPCQ. The support agent opens the contract and initiates cancellation. This single action triggers a coordinated workflow: billing is scheduled to stop on the service end date with any required proration, the OLT port is flagged for deprovisioning, the inventory system creates a recovery task for the assigned ONU (serial number pre-filled), and the contract status changes to "Pending Termination", visible to every team. If equipment is not recovered within the configured window, an automated reminder escalates to the field supervisor.

The result. Every step is connected because every module reads from the same contract record. There is no handoff email. There is no "I didn't know they cancelled." The contract is the nucleus, and every team orbits around it.

Deployment

Your cloud or ours?

ISPs handle sensitive customer data, financial records, and network credentials. Where that data lives matters, not just for compliance, but for operational confidence. Two equally legitimate deployment paths; same architecture either side.

01
Self-host on your hardware.
Deploy on your own bare metal or private cloud. You control the server, the backups, the firewall rules, and the data retention policy. Customer data never leaves your network. Many ISPs prefer this model because they already operate their own infrastructure; running one more service is natural.
02
Run as SaaS, ISPCQ-managed.
We manage the updates, backups, and security on dedicated isolated instances. Each ISP gets their own database and application instance. No multi-tenancy with anyone else, no shared resources. You still own the data and can export it at any time. This model works well for ISPs that want ISPCQ without adding another service to their infrastructure team's responsibilities, or for new operators bringing the team up to speed.
03
Migrate between modes anytime.
Same architecture either side of the boundary. Same data model, same migration path. Start as SaaS during a team transition; move to self-hosted later. Or run self-hosted and shift to SaaS during an infrastructure change. The choice is reversible; sovereignty is about that reversibility, not about picking a side once and forever.
Operating model

How teams stay aligned.

ISPCQ keeps every department on the same operational timeline: customer lifecycle, network state, financial status, and field actions.

01
Unified contract context
One record powers every workflow

One contract record powers support, billing, network, and inventory workflows. When a field technician installs equipment, the contract updates. When billing generates an invoice, the contract reflects it. When NOC detects a fault, the affected contracts are identified automatically.

  • Shared customer truth across every team
  • Faster cross-team handoffs
  • Fewer reconciliation disputes
02
Configurable workflows
Operations team adjusts; engineering doesn't ticket

Every ISP operates slightly differently. ISPCQ adapts through database-driven configuration: form fields, validation rules, approval chains, notification triggers, and regional variations. No code changes required. The operations team can adjust workflows as the business evolves.

  • Regional and operational flexibility
  • Consistent operator experience across changes
  • Safe change management without engineering bottlenecks
03
Execution visibility
Audit trails for every significant action

Every significant action is logged: who changed a plan, who approved a credit note, who closed a ticket without resolution evidence, who dispatched a technician. This creates an operational timeline that leadership can review, not for surveillance, but for process improvement and accountability.

  • Operational audit trails with before/after values
  • KPI-driven decisions
  • Accountability at scale
What this enables

Business outcomes for ISP leadership.

Less cross-system friction. Teams stop re-entering data and reconciling conflicting records between tools. A support agent doesn't need to ask "what's the billing status?" because it's on screen. A finance clerk doesn't need to email NOC asking "is this customer actually disconnected?" because the contract status is shared.

Faster operational decisions. When the COO asks "how many customers are affected by that fiber cut?", the answer is available in seconds, not after someone queries three systems and cross-references in a spreadsheet. Decision-makers see billing risk, network health, and field status in one place.

Stronger accountability. Every key action is visible and traceable across teams. When something goes wrong (a customer billed incorrectly, equipment missing after an install), the audit trail shows exactly what happened, when, and by whom. This shifts culture from "who's responsible?" to "how do we prevent this?"

Cleaner
Cross-team handoffs
Fewer
Operational disputes
Lower
Manual sync overhead
Faster
Customer-impact decisions
Clearer
Ownership of changes
Higher
Confidence at scale