One API, every client
The portal, CLI, Terraform, and MCP clients drive the same operations with the same permissions. Anything you can click, you can automate.
Virtual machines, private networks, and storage behind one product API. The portal, CLI, Terraform, and MCP clients share the same operations, permissions, and audit trail — so your team and your agents work the same way.
Everything a production cloud needs, one API away
Compute, storage, and networking provisioned in seconds — from the portal, the CLI, Terraform, or an agent. Every resource is scoped to a project, covered by quotas, and metered from the moment it exists.
Elastic compute
General-purpose and dedicated-CPU virtual machines launched from trusted images. Resize, snapshot, and rebuild through the same API that runs everything else — with root access from first boot.
Whether an engineer clicks in the portal, CI applies Terraform, or an agent calls a tool over MCP, the request lands in the same place: an operation that is authorized, audited, and metered before it runs.
$ ezgh instance create api-1 --type c2.large
✓ operation op_4t9x2c accepted
✓ audit recorded · actor mira@acme.dev via cli
✓ instance api-1 running · 14s
$ ezgh operation get op_4t9x2c
requested mira@acme.dev (cli)
changed instance api-1
audit recorded · immutable
metered project acme/productionmcp ›agent deploy-bot called instances.create
same roles · same quotas · same audit trail
Identity, quotas, audit, and metering aren’t add-ons — they’re how the platform works. Every project gets the same controls, whether it’s run by a team of engineers or a fleet of agents.
The portal, CLI, Terraform, and MCP clients drive the same operations with the same permissions. Anything you can click, you can automate.
Agents connect over MCP like any other client — scoped to a project, bounded by the same roles and quotas, and audited like any other user.
Each region runs its own control, storage, and network stack. Trouble elsewhere — even in the global control plane — does not take your workloads down.
Every operation records who asked, what was approved, and what changed — before anything runs. The trail is immutable and yours to query.
Organize work into projects with their own members, roles, quotas, and usage. Give a team — or an agent — exactly the access it needs.
Run commands and diagnostics on instances through an outbound-only agent. No public IP, no inbound SSH, no bastion to babysit.
Platform at a glance
Metered usage, predictable monthly rates, and no long-term contracts. Every project sees exactly what it used — down to the instance-hour and the gigabyte.
Virtual machines for apps, APIs, databases, and CI — provisioned in seconds.
Replicated volumes and S3-compatible object storage, attached on demand.
Private networks, public IPs, load balancers, and DNS on a dual-stack edge.
Our team can help you size a project, plan a migration, or set up access for your engineers and your agents.
Contact salesA public cloud platform: virtual machine instances, private networks, block and object storage, load balancers, and DNS — managed through one product API. The portal, CLI, Terraform, and MCP clients all share that API, so every action carries the same permissions, audit, and metering.
Agents connect through MCP, the same way your other tools do. An agent acts inside a project with the roles and quotas you grant it, and every action it takes lands in the same audit trail as everyone else’s. There is no separate, privileged side door.
Regions are autonomous failure domains with their own control, storage, and network stacks. Your instances, networks, and volumes keep serving inside their region even if another region — or the global control plane — is having a bad day.
Usage is metered per project: instances by the hour, storage per GB-month, and network resources per allocation, billed monthly with no long-term contracts. Delete a resource and its meter stops. Every project can see exactly what it used and why.
No. The managed agent connects outbound from your instance, so you can run commands, collect diagnostics, and patch systems without inbound SSH or a public address. Full root access is still yours whenever you want it.
Get started with ezghcloud
Create a project, pick a region, and go — or talk to our team about bringing your workloads over.