For CIOs, IT operations directors, and cloud architects
Investigate cost and operations drift across Azure, AWS, GCP, and StackIT.
Cloud Horizons connects billing, inventory, lifecycle, delegated tenant access, and MCP automation into an owner-ready review path. Read-only setup takes minutes; write actions stay opt-in per cloud when a team is ready to automate.
No card required. Read-only by default. Write actions opt-in per cloud.
Generated sample inventory
Resources grouped after read-only connection
| Type | Cloud | Status |
|---|---|---|
| Compute | AWS | Owner tagged |
| Kubernetes | Azure | Policy linked |
| Kubernetes | GCP | Owner tagged |
| Virtual desktop | Azure | Review needed |
| Database | AWS | Owner tagged |
MCP endpoint
$ curl https://cloud-horizons.com/api/mcp/v1 \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'
{"jsonrpc":"2.0","id":1,"result":{"tools":[...]}}Connection methods
- AWS IAM roleread-only
- Azure service principalread-only
- GCP service accountread-only
- StackIT project tokenread-only
Four clouds, your existing tags, your provider-native access model
Three steps from connected to automated
Read-only access on day one. Write actions when you flip the switch. MCP endpoint when you are ready to automate the loop.
Connect once, see everything
Read-only role per cloud. Resource inventory, tags, and billing data populate within an hour. No agents, no Terraform changes, no CRDs in your cluster.
Operate from one control plane
VM and AVD lifecycle, EA reconciliation, multi-tenant via Lighthouse and Organizations. Cost impact preview on every action. Audit log on every change.
Automate through MCP
Same surface exposed as an MCP server. Your platform team or AI agent queries inventory, triggers actions, and pulls cost data through one authenticated endpoint. Scope-limited per token.
Setup time
10 minutes
Read-only IAM role for AWS, service principal for Azure, service account for GCP. No agents, no Terraform changes.
Provider coverage
4 clouds
AWS, Azure, GCP, and StackIT with provider-native read access and no browser scraping.
Time to first useful insight
< 1 hour
After IAM access lands. Anomaly scan runs immediately, attribution and forecasts populate as the next billing snapshot arrives.
Six surfaces tied back to source data
Cost is the wedge most teams come in for. Then they find the rest. The same read-only access opens inventory, lifecycle, multi-tenant, and the MCP server for AI agents.
Resource inventory
Every VM, database, storage account, load balancer, and managed service across AWS, Azure, GCP, and StackIT. Tagged owner, age, last-touched timestamp, monthly run-rate cost. Filterable, exportable, scriptable through the API.
VM and AVD lifecycle
Schedule, snapshot, resize, deallocate, and de-provision. AVD session host pool management with cost impact preview before each action. Approval workflow for the actions that cost real money.
Multi-tenant via Lighthouse
Manage delegated subscriptions for your customers from one control plane. Per-customer scoping, per-customer tagging, per-customer billing aggregation. AWS Organizations and GCP folder IAM as the equivalents on those clouds.
EA billing reconciliation
Reconcile your enterprise agreement, MCA, or AWS Enterprise Agreement against the resources actually in use. Surface unused commitment, unconverted reservations, and spend categories that drift outside the negotiated schedule.
MCP server for AI agents
Cloud Horizons exposes inventory, costs, anomalies, forecasts, Azure APIM, and StackIT project data as MCP tools. Teams can query provider context through one authenticated endpoint with scoped tokens and audit logs.
Anomaly and cost alerts
Daily anomaly scan with hand-reviewed alerts before delivery. Per-team budget tracking, runaway resource detection, reserved instance expiry, and unconverted savings plans. Email and Slack delivery, no inbox spam.
Azure APIM monitoring
Live service state, API inventory, subscription key health, and Azure Monitor metrics (requests, latency, failures) for Azure API Management instances. Integrated into the platform workspace alongside VM and SQL monitoring.
Operations evidence
Azure APIM, App Gateway, Advisor, Service Health, Monitor, Cost Management, and Resource Graph are grouped with AWS, GCP, and StackIT cost context so each recommendation has source data behind it.
StackIT monitoring
Cost Management, project inventory, OpenStack compute instances, and S3-compatible object storage in one view. Built for EU sovereignty and DACH compliance workloads on the Schwarz Group sovereign cloud.
Integrations that already work
We read what the cloud already publishes. No agents, no sidecars, no custom CRDs in your cluster. Read-only by default, write scopes opt-in per cloud, audit logged on every action.
AWS
Cost Explorer · Compute Optimizer · Trusted Advisor · Organizations
Azure
Cost Mgmt · Advisor · Lighthouse · Entra ID · Policy
GCP
Billing API · Recommender · Asset Inventory · Folder IAM
StackIT
Cost API · Inventory · Project service accounts
Slack
Anomaly alerts · Daily digest · /cloudhorizon slash command
PagerDuty
Incident routing for spend anomalies above threshold
Opsgenie
Alternative incident routing for Atlassian shops
Terraform
Read-only state import for tagging gap detection
Webhooks
Signed POST per anomaly to your tooling of choice
API + MCP
Same surface for humans, scripts, and AI agents
Multi-tenant for managed-service partners
Lighthouse on Azure, Organizations on AWS, folder IAM on GCP. One scoped view for every customer you manage.
Deploy our published Lighthouse ARM template. Customers delegate access through the standard offer. Cloud Horizons reads delegated subscriptions the same day. Per-customer scoping, per-customer tagging, billing aggregated for your invoice run.
AWS Organizations plus AWS RAM cross-account roles cover the same ground on AWS. Folder-level IAM bindings on GCP. Project-scoped service accounts on StackIT. The partner playbook documents each one.
Start the free 14-day trial
Read-only access, all four clouds. We schedule a 30-minute onboarding call within one business day. The trial returns an owner-ready workspace you can hand to your platform team on Monday.