Settings and audit log
Set organization controls and trace activity in the audit log.
Open Settings in the sidebar's Organization group. Admins can rename the organization and change its controls. Renaming leaves the URL slug unchanged. Other members can read the name, URL and switches.
Compute and invitation controls
Kinesis serverless capacity (serverlessEnabled)
Developers can run apps on Kinesis-managed serverless CPU capacity
Developers must target an accessible grid; admins can still use serverless
Non-admin members can invite people (membersCanInvite)
Developers can invite Developers and Viewers
Only admins can send new invitations
Both settings default to enabled. Disabling serverless capacity does not stop apps already running. It prevents developers from creating, starting, restarting or rebuilding serverless apps, and from changes that need new serverless machines. Select a grid in the app's Resources settings, or ask an admin.
Disabling developer invitations leaves invitations already sent valid. Review and revoke pending invitations separately when needed.
GitHub connections
GitHub connections belong to the organization. Open Settings → GitHub or choose GitHub in the new-app wizard. Connect a GitHub account installs the Kinesis GitHub App; Link installations I can reach links an installation you can already access.
Admins can see and disconnect every connection. Developers see and use the ones they connected. A developer can step back from a connection; if they were its only connector, that removes it. A connection stays with the organization when the person who connected it leaves. Disconnecting prevents new builds and automatic deploys until the account is connected again. Existing apps keep running. Each connection shows who connected it and how many apps build through it.
The public GitHub API and MCP actions browse these connections with vcs:read and organization context. Connection management itself stays in the portal.
Single sign-on
Contact support to connect your organization’s identity provider. Once enabled, this Settings section shows its verified domains and sign-in link. Admins choose the first-sign-in role, open the provider setup portal, refresh domains and optionally require SSO. The requirement uses a 24-hour browser sign-in window and does not revoke existing API keys or connected apps. Read the SSO guide for setup, bringing existing members over, recovery and offboarding.
Admins also register the organization's SSH public keys in Settings. Assignment depends on server eligibility; use the server setup reference.
Your personal profile, sign-in methods, API keys and connected apps are in Account settings, not organization Settings.
Read the audit log
Admins open Organization → Members and scroll to the audit log to see who changed the organization's members, settings, projects, apps, grids, servers and GitHub connections. Searching on Members also filters the log to entries by or about the matching people. The latest entries also appear under Recent activity on Overview.
The log retains one year of activity, including recorded platform events. SSO setup, domain, default-role and requirement changes, SSO joins, owner recovery sign-ins and resource moves between organizations are recorded. Actions Kinesis support takes while signed in as a member are marked as support; in organizations that require SSO, each such support sign-in is recorded too. Expand a row with a recorded diff to inspect field changes.
The API returns entries in latest-recorded-first order. Follow nextCursor for older entries. Each entry includes:
actor
Person who acted; null for the platform or an account that is gone
action
Recorded operation, such as memberInvite, appCreate or serverProvision
outcome, errorCode
Success or error, and a code for a failed action
details
Action-specific context
changes
Field paths, change kinds and before/after values where recorded
at, role, viaSupport
Time, actor's role and whether support acted while signed in as the person
Entries also carry id, resourceType, resourceId, resourceName and message. Secret changes are marked as redacted; their values are not retained in the diff.
API and MCP
GET /org?org=acme / get_org reads settings, caller role and frozen state. PATCH /org?org=acme / update_org changes the name or compute/invitation switches and requires org:write plus an admin role. SSO controls are portal-managed and are not fields on this public operation.
GET /org/audit?org=acme / list_audit_log requires the audit:read scope (or a full-access key), plus an admin role. org:write and members:read do not imply audit access, and keys limited to projects can't read the log. The audit log names members, so grant this scope deliberately.
In compact MCP mode these actions live under account and account_write. Older connected apps must be re-approved with the new scopes before they can use them. See OAuth for MCP clients.
Was this helpful?
