How it works

From payload to delivery report in under 500ms

A deep and honest look at every layer of the platform. No "magic" — just solid engineering, battle-tested components in production, and explainable design decisions.

Architecture

End-to-end non-blocking pipeline

Every layer is engineered to sustain billions of messages/month — with guaranteed delivery, smart retries, distributed rate-limiting, and full per-message observability.

Your Backend

App / CRM / ERP

API Gateway

REST · Webhooks · SDKs

Priority Queue

Per workspace and type

Parallel Workers

Non-blocking pipeline

Meta Cloud API

WhatsApp Business

Webhooks

Callbacks · Status

Throughput

100% non-blocking pipeline. Queues prioritized by tenant and message type.

Resilience

Smart retries with exponential backoff and periodic reconciliation with the Meta Cloud API.

Observability

Every message traceable end-to-end — logs, metrics, tracing, and alerts.

Message flow

8 steps from backend to WhatsApp

Step 01

Ingestion

POST /v1/messages arrives at the edge

The gateway validates schema, token signature, dedup by idempotency_key, and routes to the appropriate queue. 100% non-blocking — we return 202 in under 40ms with a persisted message_id.

API GatewayJWT + scopesNative 24h idempotencyOpenAPI 3.1 validation

Step 02

Queueing

Priority queue by workspace + type

Messages are routed to separate queues by priority (critical, high, default, low) and workspace. This ensures heavy marketing volume never delays an OTP. Distributed rate-limiting per template and number.

Priority queue per workspaceFour levels (critical/high/default/low)Distributed rate-limiting

Step 03

Processing

Parallel workers consuming in a loop

Each worker grabs a batch, hydrates template variables, applies segmentation rules, and fires to the Meta Cloud API. Auto-scaling pool with backpressure — we never starve the downstream API even at peak.

Auto-scaling parallel workersAdaptive backpressureConnection pool with Meta

Step 04

Meta Cloud API

Delivery to WhatsApp Business

Meta returns an id, then notifies status (queued → sent → delivered → read). On error, we classify as retryable / terminal and handle with the appropriate policy.

Meta Graph API v19+Aggressive timeout (2s)Circuit breaker per number

Step 05

Smart Retries + Reconciliation

Failure is not the end — it's the beginning of reconciliation

Retryable errors go to a retry buffer with exponential backoff (2s → 4s → 8s → 16s → 32s → 1min → 5min → 30min). After 7 attempts, the message enters continuous reconciliation with Meta every 6h for 7 days.

Retry bufferExponential backoffAutomatic reconciliation

Step 06

Persistence

Data core as source of truth

Every state transition is persisted with timestamp, Meta metadata, billable cost, and tracing id. Composite indices by workspace + status + created_at enable queries in milliseconds even with billions of records.

Physical isolation per workspaceAutomatic TTL for logsComposite indices

Step 07

Outbound webhooks

Your backend receives every event

Each transition generates a webhook signed by HMAC SHA-256 to your endpoint. Automatic redelivery with backoff up to 24h, ordering by message_id, and manual replay via dashboard for the last 30 days.

HMAC SHA-256Redelivery workerReplay dashboard

Step 08

Observability

Every message traceable end-to-end

The entire pipeline produces structured logs, metrics (Prometheus), and tracing spans (OpenTelemetry). A single trace_id takes you from ingestion to outbound webhook. Public Grafana dashboards per tenant.

OpenTelemetryPrometheus + GrafanaStructured JSON logs
Pillars

Three architectural decisions that sustain everything

Native multi-tenancy

Each workspace has its own queues, rate-limits, and physically isolated storage. Real isolation — not logical filter that can leak.

Layered security

Scoped tokens, zero-downtime rotation, PII encrypted at rest (AES-256) and in transit (TLS 1.3), immutable audit log.

Total observability

A single trace_id takes you from the client API to the outbound webhook. Public dashboards per tenant. No secrets about what happens inside the platform.

Operational guarantees

What CCX commits to in writing

Beyond SLA, a list of commitments that go into Enterprise contracts and work by default on other plans.

  • Idempotency guaranteed for 7 days on every endpoint
  • No message sent twice due to internal failure
  • Retry buffer with up to 7 attempts + 7 days of automatic reconciliation
  • Webhooks with 24h automatic redelivery and manual replay
  • Public SLA with financial credits on breach
  • Real-time status page and public postmortems <48h
  • LGPD, GDPR, SOC 2 Type II auditable with DPA ready
  • Full data export anytime (CSV, JSONL, Parquet)

Read this far? You're ready to test.

Spin up a sandbox in 2 minutes and send your first approved template. No card, no lock-in.