WooSunLab
INFRASTRUCTURE

Where we use the cloud

WooSunLab's diagnostic and observation features are running on real servers today. This page shows what is live right now — measured the moment you open it — and what infrastructure the pilot and GA phases will add.

Live right now

These are real services you can call from this site today. Status and response times below are measured live as you view this page.

Website inspection (inspect)

Checking…

Search visibility observation (observe)

Checking…

AI answer observation (ai-observe)

Checking…

Analytics summary (analyze-summary)

Checking…

Rate limits and 24-hour caching protect the demo. · measuring…

What it runs

WooSunLab infrastructure workloads and status
Area What it does Status
Audit workers Rule-based inspection of public website signals (structure, metadata, content) Live
Search & AI observation Queries search engines and AI answers to check how a business appears Live
Cache & rate guards 24-hour observation cache, per-IP rate limits protecting the service Live
Analytics summary AI summaries over aggregated numbers only — raw data never leaves the browser Live
Report generation Customer reports combining diagnosis and observation Pilot
Evidence store (Trust Log) Audit trail of evidence, approvals, and executions Design
Tenant isolation Per-customer data separation for pilot customers Design
Observation scheduler Recurring observation and change-detection queues Planned

Agent workflow architecture

For each pilot customer, the platform assembles an agent workflow. Observation and drafts are automated; every outbound execution passes a human approval gate. This is assembled per engagement, after a pilot consultation.

Observation workers Live Job queue Pilot LLM gateway Live Drafts & proposals ★ APPROVAL GATE — the owner decides ★ No execution without approval Executor Pilot Publish · Change · Send Trust Log — evidence, approvals, executions Pilot
No execution without approval — this gate is the center of the architecture.
Live Pilot · Design

Cloud-agnostic by design — deployable onto managed services from any major cloud:

Cloud mapping per architecture component (illustrative)
Component Azure AWS Google Cloud
Observation & audit workers Container Apps Fargate (ECS) Cloud Run
Job queue & scheduling Service Bus / Functions Timer SQS / EventBridge Scheduler Pub/Sub / Cloud Scheduler
LLM gateway Azure OpenAI Service Amazon Bedrock Vertex AI
Approval gate app App Service App Runner Cloud Run
Evidence & log store (Trust Log) Cosmos DB + Blob Storage DynamoDB + S3 Firestore + Cloud Storage
Observability & alerts Azure Monitor CloudWatch Cloud Monitoring

Custom agent onboarding

Need an agent workflow shaped around your operations? We scope it in a pilot consultation, then assemble it on the platform.

Talk to us

Principles

Zero-storage first

The demo stores nothing server-side; from the pilot phase we keep only the minimum, isolated per tenant.

Every external call guarded

Search and AI calls are wrapped in caching and rate limits for cost and quality control.

Vendor-neutral

Containers and standard APIs first, no single-cloud lock-in.

Built with

Astro React Tailwind CSS Serverless Functions Search & AI observation APIs