There is considerable debate in the REST community whether or not transaction support is needed and possible. This chapter’s contribution to this debate is threefold: we define a business case for transactions in REST based on the Try-Cancel/Confirm (TCC) pattern; we outline a very light-weight protocol that guarantees atomicity and recovery over distributed REST resources; and we discuss the inherent theoretical limitations of our approach. Our TCC for REST approach minimizes the assumptions made on the individual services that can be part of a transaction and does not require any extension to the HTTP protocol. A very simple but realistic example helps to illustrate the applicability of the approach.