The Automation Factory

n8n Workers (x120 RPA)

The brain of the automation factory. Each worker executes workflows independently with its own dedicated resources. Kubernetes auto-scaling ensures you always have enough workers for current load.

• •Worker pool managed by n8n Main + Redis queue
• •Each worker: 0.5 CPU core, 512 MB RAM baseline
• •Scales from 10 (idle) to 120+ (peak load)
• •Built-in retry logic and error recovery

n8n Main + Redis (MCP Server)

Central orchestrator that receives workflow triggers, queues execution tasks in Redis, and distributes work to available workers. Provides the web UI for workflow design.

• •Redis acts as distributed job queue
• •Web UI for visual workflow builder
• •Persistent execution history and logs
• •Webhook endpoints for external triggers

LLM (Claude / GPT-4)

Large language models integrated via MCP (Model Context Protocol) for agentic workflows. Enables AI-powered decision making, content generation, and intelligent automation.

• •Persistent tool memory (no context resets on R640)
• •Multi-step reasoning and planning
• •Integration with n8n workflows via API
• •Custom prompts and fine-tuned responses

Gotenberg (PDF Microservice)

Headless Chrome-based PDF generation service. Converts HTML templates to professional PDFs with full CSS support, custom fonts, and server-side rendering.

• •Scales independently from n8n workers
• •Handles invoices, reports, certificates
• •R640: 2500+ PDFs/hr with burst capacity
• •Custom headers, footers, and watermarks

VAPI (Voice AI - PSTN / WebRTC)

Voice AI integration for phone-based automation. Handles inbound/outbound calls with natural language understanding, speech-to-text, and text-to-speech synthesis.

• •150+ concurrent calls on R640 cluster
• •Real-time transcription and sentiment analysis
• •Connects to n8n workflows via webhooks
• •Multi-language support and custom voices

Caddy (Low Latency / SSL Ingress)

Automatic HTTPS reverse proxy with instant SSL certificate provisioning. Routes external traffic to internal services with minimal latency and built-in security.

• •Automatic Let's Encrypt SSL certificates
• •HTTP/2 and HTTP/3 (QUIC) support
• •Load balancing across worker pods
• •DDoS protection and rate limiting

Terraform (Infrastructure as Code)

All infrastructure defined in code and version-controlled. Reproducible deployments across dev, staging, and production with automated rollbacks and drift detection.

Grafana (Network Latency / Logs)

Real-time monitoring dashboards with metrics from all cluster components. Network latency graphs, error rate tracking, and centralized log aggregation from every pod.

Resend (The Messenger)

Modern email delivery API for transactional emails. Handles workflow notifications, automated reports, and customer communications with high deliverability.

Vercel Dashboard (Reporting)

Centralized dashboard for cluster health, deployment status, and analytics. Accessible from any browser with role-based access control.

Supabase (Private DB / User Mgmt)

Self-hosted Supabase provides PostgreSQL databases, authentication, real-time subscriptions, and storage APIs. All running within your cluster—no external SaaS dependencies.

Redis + Persistence

In-memory data store for job queues, caching, and session management. Persistence enabled to survive cluster restarts without losing queued workflows or cache data.