Skip to main content
The CLI, the REST API and the MCP tools identify a resource by its ID and name, and do not return a portal link with it. Every portal page for those resources has a URL you can build from those values, so a script or an AI assistant can hand a teammate a link to exactly what it found. Links on this page use https://platform.ankra.app. If you sign in to Ankra at a different address, use that one instead. Opening a link needs a signed-in session in the organisation that owns the resource.

​
Clusters

Every cluster lives under the same path, however it was created:
The imported segment is a historical name, not a cluster type. A cluster Ankra provisioned on UpCloud, Hetzner, OVH, DigitalOcean, AWS or any other provider uses the same path as a cluster you imported. There is no page at .../imported/<cluster-id> on its own: overview is the cluster’s landing page. <cluster-id> is the cluster’s id, as returned by ankra cluster list -o json and the MCP get_cluster_details tool. Other cluster pages replace overview with:

​
Stacks, manifests and add-ons in a stack

A deployed stack opens in the stack editor:
To open one manifest or add-on inside the stack, add the activeTab and activeSubTab query parameters:
<stack-name>, <manifest-name> and <addon-name> are the names ankra cluster stacks list and the MCP list_stacks and get_stack_details tools return. URL-encode a name that contains anything other than letters, digits, -, _ and .. The : after manifest or addon can be written as %3A. A link copied from the browser also carries clusterId and stackName query parameters. They repeat values that are already in the path, so you can leave them out. These links need a deployed stack. A stack that exists only as a draft has no page under its name; open it from the cluster’s Stacks list instead.

​
An add-on’s own page

Each add-on on a cluster also has an overview page outside the stack editor:
<addon-name> is the name ankra cluster addons list shows.

​
Operations

An operation (an execution) on a cluster opens at:
Take both values from ankra cluster operations list -o json: <execution-id> is the execution’s id and <cluster-id> is its cluster_id. There is no page at .../operations/<execution-id> without /kanban. An execution whose scope is platform also has an organisation-level page, which is the only one available when it has no cluster_id:
That page opens platform executions only. A cluster execution is not found there, so use the cluster link above for it.

​
Pipeline runs

A pipeline run opens at:
Take both values from ankra pipeline get <run> -o json: <repository-id> is repository_id and <pipeline-run-id> is id. Do not use the run’s run_id field, which identifies a different record. The Open this run in Ankra link in a pull request’s pipeline status comment points at the same page.