Skip to main content
This reference is generated from the spend control OpenAPI specification. Each endpoint page shows the request, parameters, response schema, and live examples.

​
Base URL

​
Two credentials

An agent key can never write policy, and a dashboard token can never authorize an action as an agent. See authentication.

​
Response envelope

Single resources wrap the object in data:
The two authorize endpoints add the decision alongside it, and a cosignature when one was issued:
The cosignature key is absent, not null, when no signature was issued.
Control-plane responses carry no livemode and no request_id. The ledger API includes both. If you are building one client for both, do not assume a shared envelope.

​
Pagination

Workspace lists take limit (default 50, clamped 1..200) and starting_after, which is the id of the last record on the previous page rather than an opaque cursor.
Three things differ from the ledger’s pagination:
  • The parameter is starting_after, not cursor.
  • has_more is true whenever a page filled to the limit, so a final page that exactly fills it still reports true. Fetch until you get an empty page rather than trusting the flag alone.
  • An unknown or foreign starting_after is ignored and returns the first page rather than an error.
GET /control/v1/workspaces and GET /control/v1/workspaces/{slug}/funds are not paginated. Most lists are newest first. Memberships, invitations and policy modules are ascending, and policy modules paginate by name rather than by id.

​
Idempotency

Keys are scoped per agent. A replay returns the original intent and the original decision, at the original status.

​
Errors

See the errors reference for every code, and for the rules that can appear in decision.rule.

​
Next steps

Authorize an agent action

The five-minute path through these endpoints.

Ledger API reference

The other API on this host.