Open Source · Self-Hosted · Enterprise-Ready

Cloud-native self-service
on your infrastructure.

Stop being a third-class citizen in your own data centre. Cloudyard gives developers, platform teams and data scientists the self-service experience they demand — without a single ticket, without AWS, without your data leaving your rack.

Get Early Access See What's Included
portal.cloudyard.io — Service Catalogue
CatalogueCreatePipelines
my-company · platform
Platform
⊞  Catalogue 23
⚡  New Service
◈  Deployments
📊  Dashboards
Data Platform
⛁  Data Lake
⟳  Pipelines
◉  Workspaces
🔍  DataHub
Infrastructure
☸  Clusters
🔑  Secrets / PKI
Service Catalogue
23 components across 4 teams
⌕Search services, APIs, data sets...⌘K
analytics-apiService
FastAPI + Postgres. GitLab CI auto-deploy.
data-team
ml-pipeline-churnData
Airflow DAG · MinIO · DataHub lineage
ml-team
talos-prod-01Infra
6-node Talos · Cilium · Proxmox
platform
ds-workspace-alexWorkspace
Coder · PyCharm · A100 GPU · pip self-service
data-science
internal-chatAI
OpenWebUI · private LLM · OIDC-gated
platform
minio-s3-clusterStorage
12TB MinIO · S3-compatible · distributed
platform
// The Problem We're Solving

Infrastructure that treats developers
and data scientists as first-class citizens.

You shouldn't need to raise a ticket to spin up a notebook environment, access a data bucket, or install a Python package. Cloudyard fixes the self-service gap that exists in every traditional data centre and on-prem setup.

🎫 The Ticket Tax

Every infrastructure request goes through a queue. Want a new namespace? Ticket. Need S3 access for your pipeline? Ticket. Want to install PyTorch? You guessed it. Cloudyard eliminates the queue through OIDC-backed self-service and GitLab CI-driven automation.

🧪 Data Science in the Dark

Data scientists are the most expensive to provision for and the most commonly ignored. No GPU access. No pip freedom. No data lake visibility. Cloudyard gives them full Coder workspaces with PyCharm and VSCode, self-service package management, and transparent OIDC access to every data asset they're authorised to see.

🔓 SSO That Actually Works

In most on-prem setups, identity is an afterthought. Separate credentials for every service. Cloudyard is OIDC-native — every service from the dev portal to the data lake is wired to the same identity provider from day one.

💸 The VMware Exit

Broadcom's acquisition of VMware changed the economics for thousands of companies. Cloudyard gives you a clean, modern starting point — Talos Kubernetes, no proprietary licensing, no vendor lock-in. Keep what you know, replace what hurt.

// Platform Capabilities

Everything pre-wired.
Nothing left to configure.

Cloudyard isn't a list of components — it's a set of battle-tested integrations. Each piece knows about the others from the moment it deploys.

  • ✓
    Backstage service catalogue — Every service your team deploys lives here. Discoverable, documented, observable. New engineers are productive in hours.
  • ✓
    GitLab CI scaffolding — Click a template in the portal, a pipeline runs, your service is deployed. No YAML hand-coding required.
  • ✓
    Kubernetes namespaces on demand — Claim a namespace, get RBAC and network policies pre-configured. No platform team ticket required.
  • ✓
    PKI and secrets built in — Internal CA, cert-manager integration. Your services get TLS with no manual certificate management.
⚡Create New Service
🐍 FastAPI Microservice
Python · Postgres · Prometheus metrics · Dockerfile included
Selected
📊 Airflow Data Pipeline
MinIO source · DataHub lineage · Slack alerts
☸ Kubernetes Workload
Deployment · Service · HPA · Ingress pre-configured
Deploy via GitLab CI →
  • ✓
    Coder workspaces — Full cloud development environments running on your Kubernetes cluster. Provisioned in under 2 minutes.
  • ✓
    PyCharm & VSCode — Data scientists and engineers get their preferred IDE, browser-based or via SSH.
  • ✓
    pip and npm freedom — Package management that just works. No tickets, no blocklists, no waiting for IT approval.
  • ✓
    GPU access on demand — NVIDIA GPUs exposed as workspace resources. Claim what you need, release when done.
◉Coder Workspaces — my-company
3 active
Py
alex-datascience-ws
8 vCPU · 32GB RAM · 1× NVIDIA A100 · Python 3.11
RunningOpen →
VS
sarah-backend-dev
4 vCPU · 16GB RAM · Node 20 + Python 3.12
RunningOpen →
Jn
james-notebook-ml
4 vCPU · 16GB RAM · JupyterLab
Stopped
  • ✓
    MinIO S3-compatible data lake — Scalable distributed object storage on your own hardware. Fully compatible with any S3 SDK.
  • ✓
    OIDC access, zero manual IAM — Data lake access is tied directly to your identity provider. Transparent, auditable, no tickets.
  • ✓
    DataHub data catalogue — Automated lineage, schema discovery, and data quality tracking built in.
  • ✓
    Isolated by default — Each team gets their own bucket namespace. Cross-team sharing is explicit and audited.
⛁ Data Lake — my-companyOIDC • All access audited
⊡ml-training-data2.4TByour team
⊡raw-events-2024890GBshared · read
⊡model-artefacts340GByour team
// Your access via OIDC — [email protected]
✓S3 keys auto-rotated via OIDC token exchange
✓Access audited in Grafana Loki
✓No manual IAM configuration required
  • ✓
    Airflow pre-wired to MinIO — DAGs can read and write to your data lake from day one.
  • ✓
    Argo Workflows for compute tasks — ML training, batch processing. Runs on your Kubernetes cluster with full resource governance.
  • ✓
    DataHub lineage automatic — Every pipeline step emits lineage metadata. You always know where data came from.
  • ✓
    Scaffold from the portal — Click "New Pipeline" in Backstage. GitLab CI creates your boilerplate, wires it to the data lake, pushes to your repo.
⟳ ml-pipeline-churn — last run 4m ago · success
⛁
MinIO
Source
→
🔧
Feature
Eng.
→
🧠
Model
Train
→
📊
Eval &
Log
→
⛁
MinIO
Output
14m 22s
Avg. runtime
98.4%
30d success rate
2.1GB
Data processed/run
// How It Works

From bare metal to running platform
in a single day.

01

Point at your infrastructure

Run the Cloudyard bootstrap. Point it at Proxmox, VMware or OpenStack. It provisions Talos Kubernetes automatically.

02

Choose your stack

Select from the Cloudyard component library. Dev portal, data lake, workspaces, observability. Pick what you need.

03

Wire your identity

Connect Keycloak, Auth0 or Azure AD. Every deployed service inherits identity automatically.

04

Hand it to your team

Developers get the Backstage portal. Data scientists get Coder workspaces. Everyone gets what they need, immediately.

// Pricing

Free to run forever.
Subscription for the enterprise layer.

The Cloudyard platform is fully open source with no seat limits. Subscribe to unlock enterprise SSO, premium templates, dashboards and SLA-backed support.

// Community
Open Source
£0
forever, no limits
The full platform. Every component, every integration. Community-supported.
  • ✓Backstage developer portal
  • ✓Talos Kubernetes bootstrap
  • ✓MinIO data lake
  • ✓Coder workspaces
  • ✓Airflow + Argo Workflows
  • —SSO / OIDC integration
  • —Premium Coder templates
// Enterprise
Cloudyard+
Contact
per deployment / month
Everything in Community plus enterprise SSO, premium scaffolding and SLA-backed integration support.
  • ✓Everything in Community
  • ✓SSO + OIDC configuration
  • ✓Alertmanager configs + rules
  • ✓Premium Coder DS templates
  • ✓Data pipeline scaffolding
  • ✓SLA integration support
// Via DC Partner
DC Tenants
Ask DC
bundled with hosting
Running inside a Cloudyard partner data centre? You may already have access.
  • ✓Full Cloudyard+ features
  • ✓Managed by your DC provider
  • ✓Your data stays in their rack
  • ✓Flat monthly billing via DC
Get Early Access

Your infrastructure deserves
a cloud-native experience.

Join the early access list. We'll reach out when your slot is ready.

$ cloudyard bootstrap --provider proxmox --components all