About Anthropic Administrator
The Anthropic Admin API allows programmatic management of organizational resources, including members, workspaces, and API keys.
Describe what should happen in Anthropic Administrator, and when, in plain words. Whenever builds the workflow and runs it on a schedule, on a webhook, or when another app sends an event.
Connection requirements, provider permissions and optional AI usage apply. Read the runtime limits.
The Anthropic Admin API allows programmatic management of organizational resources, including members, workspaces, and API keys.
Through its API, using the credentials and permissions required by the provider.
Connect Claude, Cursor or Codex to the Whenever MCP server and your coding agent can build and deploy the Anthropic Administrator workflow. It runs on Whenever afterwards, on its schedule or webhook, with the chat closed.
Set up the Whenever MCP serverThe Anthropic Admin API is for the people who run a Claude API organization, not for calling models. It manages who belongs to the organization, which workspaces exist, who can use each workspace, and the status of API keys, and it reports usage and cost. Automating it typically means turning recurring admin chores into checks: a weekly list of keys nobody has used, a monthly cost breakdown per workspace, or an onboarding step that invites a new hire with the right role.
Authentication normally uses an Admin API key, which starts with sk-ant-admin and is different from an ordinary API key; an OAuth token with the org:admin scope also works. Only organization members with the admin role can provision an Admin API key, and the Admin API is unavailable for individual accounts; an organization has to be set up first. Treat this credential as highly privileged, since it can remove members and disable keys across every workspace.
The resources follow the organization structure. Users can be listed, have their role updated or be removed. Invites are created with an email and role, and expire after 21 days if not accepted. Workspaces can be created, renamed and archived, and workspace members carry their own workspace role. API keys can be listed, filtered by workspace or status, renamed, and set to inactive or archived; new API keys are created in the Console, not through the Admin API. The usage report returns token counts in time buckets that can be grouped by workspace, model or API key, and the cost report returns daily spend that can be grouped by workspace.
Because these calls change access, write disabling and removal as explicit rules with a list of exceptions, and have the workflow report what it would change before you let it act. Cost figures in the reports are the organization's own billing data, so they suit chargeback sheets where numbers must be exact.
Authored starter briefs to configure with your accounts, not preconfigured workflows.
Uses the cost report grouped by workspace to build a chargeback sheet.
A report-only check: it lists keys for a person to review and disables nothing.
Turns a row in an onboarding sheet into an organization invite with a fixed role.
An organization admin creates it in the Claude Console. Ordinary API keys cannot call Admin endpoints, and individual accounts without an organization cannot use the Admin API.
No. It can list keys, rename them and change their status, for example to inactive. Creating keys remains a Console action.
Yes. The cost report returns spend grouped by workspace, and the usage report returns token counts that can be grouped by workspace, model or API key.
It depends on the key type. Anthropic documents that personal keys stop working when their user is removed, while workspace API keys keep working. Check key ownership before an offboarding workflow removes someone.
Only with explicit rules, an exceptions list and a report-first stage. The Admin API key can change access across the whole organization, so store it carefully and limit who can edit the workflow.
Describe your first workflow in plain words. One stays live for free, no credit card.
No credit card required