From Pilot to Production: How Mediterranean Algae Scaled on AWS with €8,600 in PoC Funding
The Client
Mediterranean Algae (MEDALGAE) is a Spanish climate infrastructure company based in Alicante, focused on coastal resilience and marine water regeneration. Its technology line combines POSINODE®, an industrial telemetry station for automated water-quality monitoring; SEANGULAR®, a SaaS platform for real-time water-quality data, alerts and reporting; and BIOREMED ALGAE®, a modular macroalgae-based bioremediation system for ports, marinas and aquaculture facilities.
- Industry:Climate Technology Product Manufacturing
- Company Size:11-50
- Country:Spain
Cloudvisor helped us move Seangular, our real-time water-quality data platform, from a serverless pilot stack to a production environment on AWS in about seven weeks. Their engineers adapted the architecture as our requirements evolved and left us with infrastructure as code and automated deployments that our own team now runs. That gave us a solid base to onboard our first customers and to start working on our data and AI roadmap.
Challenges
As Seangular moved from pilot to commercial launch, Mediterranean Algae needed infrastructure that could support its first customers:
- Production readiness – The pilot ran as a single staging setup, and going live with customers required separate dev, staging and production environments.
- Multi-provider stack – The application, database, message broker and cache ran on four different serverless providers (Vercel, Neon, CloudAMQP and Upstash), which suited the pilot but needed to come together on one platform for production.
- Automation, security & visibility – Moving to production called for automated deployments, centralised monitoring and an enterprise-grade security baseline.
- Industrial IoT connectivity – Industrial PLCs in the field needed to stream data to the MQTT broker securely over TLS, without a VPN on every physical device.
With customers ready to onboard, the team wanted a single, production-grade AWS foundation that its own engineers could run and build on, including its upcoming data and AI work.
Solutions
To solve these challenges, Cloudvisor built a production-grade AWS platform for Seangular across three environments (dev, staging, production), adapting the architecture as requirements evolved during delivery.
Phase 1: AWS Foundation & Security Baseline
- Provisioned three separate AWS accounts, each with a 3-tier, 2-AZ VPC (public, private-app, private-database), using Terraform infrastructure-as-code.
- Deployed a security baseline—CloudTrail, GuardDuty, Inspector, and IAM with MFA enforcement—with Secrets Manager and SSM Parameter Store for credentials and configuration.
Phase 2: Containerized Application Platform
- Hosted both the FastAPI backend and React frontend on Amazon ECS Fargate behind an ALB, with CloudFront host-based routing and AWS WAF.
- React frontend on ECS Fargate, keeping the full stack in the Spain region (eu-south-2) for data residency.
Phase 3: IoT Messaging & Data Layer
- Configured Amazon MQ for public MQTT access over TLS so PLCs connect directly without VPN, using Virtual Topics so each message is processed by exactly one API instance.
- Aurora PostgreSQL in dev/staging, and a production database on self-managed PostgreSQL 18 with TimescaleDB on Graviton EC2, for full time-series support.
Phase 4: CI/CD Automation & Monitoring
- Automated deployments for the API and web app via GitHub Actions with OIDC authentication, building Docker images to ECR and deploying to ECS per environment.
- Set up CloudWatch alarms with SNS alerting and budget alarms wired to Slack and email.
AWS Services Used
- Amazon ECS (Fargate)
- Amazon MQ
- Amazon Aurora PostgreSQL
- Amazon EC2 (Graviton)
- Amazon ElastiCache for Redis
- Amazon CloudFront & AWS WAF
- Application Load Balancer (ALB)
- Amazon ECR
- AWS Secrets Manager
- Amazon GuardDuty
- Amazon CloudWatch
- AWS CloudTrail
Results
Mediterranean Algae now runs Seangular on a unified, production-ready AWS platform, with automated deployments and a solid base to onboard its first customers.
Production-ready in ~7 weeksKick-off on June 1st and production live by mid-July 2026.
Supported by AWS fundingCo-funded through an AWS Proof of Concept program, with €8,600 in approved AWS funding.
Four providers, one platform Vercel, Neon, CloudAMQP and Upstash consolidated into a single AWS environment.
Automated, codified deployments Three environments fully defined in Terraform, with GitHub Actions CI/CD and OIDC.
Security from day oneGuardDuty, Inspector, CloudTrail, WAF and MFA enforced across all three AWS accounts.
Full visibilityCloudWatch alarms and budget alerts sent to Slack and email.
Real-time data, end to end Dashboards update in real time over WebSockets, with PLCs streaming over MQTT/TLS.
