Platform team Northwind
prod-east cluster ••••
12 NODES
“Three internal tools replaced with one workflow surface.”
Clusters Workflows Audit
Provision, operate, and scale Kubernetes without the overhead.
The only control plane built to help platform teams ship production clusters with confidence.
Runs on SOC 2 Type II & ISO 27001 controlled
infrastructure, with every workflow audited and encrypted.
api.usebey.com LIVE
$ curl -X POST \
https://api.usebey.com/v1/kubernetes/clusters/talos \
-H "Authorization: Bearer $BC_TOKEN" \
-d '{"name": "prod-east", "region": "nbg1"}'
{
"cluster": "cls_8f2a",
"workflow": "wf_talos_01J",
"status": "PROVISIONING",
"steps": 12,
"eta_seconds": 238
} Platform teams at
- Northwind
- Vector Labs
- Ledgerloop
- Relay/HQ
- Meridian
- Atlasworks
- Prism
- Halcyon
ML infrastructure Vector Labs
gpu-training workflow ••••
47 RUNS / DAY
“GPU pools and recovery, same control plane as our clusters.”
Clusters Workflows Audit
Fintech ops Ledgerloop
payments-eu org ••••
99.99 PERCENT UPTIME
“One audit trail for compliance, one runbook for on-call.”
Clusters Workflows Audit
Gaming multiplayer PRISM
match-usw2 region ••••
3.8 MIN PROVISION
“Regional clusters up in under four minutes, on demand.”
Clusters Workflows Audit
Running thousands of clusters a month.
10k+ clusters under management
99.99% provisioning success rate
< 4 min average provision time
40+ regions supported
Built from the ground up for platform teams.
Beyond Cloud is not a dashboard bolted onto someone else’s API. The runtime, the workflow engine, and the operator surface were designed as one system so the work of running Kubernetes feels like running a product.
Close to the control plane.
Provisioning, recovery, and state all run inside the same Go binary that serves your API. No cold starts, no queue lag between tiers, no arguments about which service owns cluster state.
When a workflow step fails, the recovery loop can see it in under a second and react before your on-call rotation notices.
Scale across every cluster.
The same primitives handle a lone staging cluster and a fleet of regional production control planes. Orgs, secrets, and audit trails are shaped for multi-team ownership from day one.
Grow from one cluster to hundreds without rewriting the way your platform team thinks about work.
Flexible building blocks built on Kubernetes.
Beyond Cloud exposes the pieces you already think about every day — clusters, workflows, secrets, identity — as first-class, composable primitives. Assemble an internal platform that fits your team instead of bending your team around someone else’s product.
Your team
RequestsPolicyAudit
Platform API
HTTPWebSocketJWT
Beyond Cloud
Workflow EngineStep HistoryRecovery
Cloud providers
HetznerSnapshotsNetwork
Control planes
TalosNodesKubeconfig
CLUSTER OPS
Cluster Ops
Every primitive a platform team needs to stand up, scale, and evolve production Kubernetes.
prod-east HEALTHY
n-01
n-02
n-03
n-04
n-05
n-06
n-07
n-08
n-09
n-10
n-11
n-12
- Provisioning Talos-based cluster creation with a twelve-step, restartable workflow.
- Scaling Grow or shrink pools without drifting between dashboard and reality.
- Upgrades Rolling control-plane and node upgrades coordinated by the workflow engine.
- Node pools Dedicated pools for GPU, spot, and workload-specific node types.
- Networking Hetzner private networks and firewalls wired at provision time.
- Kubeconfig Scoped kubeconfig issuance tied to org and user identity.
WORKFLOWS
Workflows
A durable execution surface that treats infrastructure work like any other shipped product.
talos-provision-2419 7 / 12
- hetzner.snapshot.resolve
- network.private.create
- servers.provision
- firewall.attach
- cloudinit.apply
- talos.machineconfig.gen
- talos.apply
- talos.bootstrap
- kubeconfig.issue
- relationships.link
- health.verify
- notify.complete
- Engine Event-driven job queue with step history persisted to SQLite.
- Recovery Stuck or interrupted workflows resume from their last good step.
- Scheduling Cron and event-driven triggers for recurring platform jobs.
- Secrets AES-256-GCM encrypted credentials scoped to a workflow run.
- Retries Exponential backoff and per-step retry policy out of the box.
- Cancellation Cancel any in-flight workflow without leaving orphaned resources.
OBSERVABILITY
Observability
The picture on the dashboard matches the picture in production, because the source of truth is shared.
system:health LIVE
- 200
cluster.provision.complete - 201
workflow.step.advance - 429
hetzner.rate.limit - 200
kubeconfig.issue
- Metrics Prometheus-compatible endpoint for control plane and workflow health.
- Logs Per-workflow log stream with structured fields and step correlation.
- Audit Every cluster and org mutation recorded with actor and reason.
- Entity graph Live view of clusters, workflows, and resources as a single graph.
- Live status WebSocket updates for provisioning, health, and system events.
- Alerts Route workflow failures and threshold breaches into your existing paging.
Start shipping today.
Stand up your first cluster in minutes, or walk through the platform with our team.