Crinkl didn't begin with the ambition of building an “Internet Labor Market.” The initial goal was much more practical: making a receipt-rewards system actually work in the messy reality of commerce data. That meant handling inconsistent receipts, imperfect OCR, and the need to give users immediate feedback when they scan while still producing data reliable enough to justify paying rewards. Early versions therefore lived inside a monolithic system focused on a simple end-to-end loop: users scan receipts, the system interprets them, and rewards are issued.
Only after that system worked did the team begin decomposing it into its underlying primitives. Receipt scanning for incentives is already a proven behavior used by tens of millions of people on centralized platforms where brands fund experiments to measure incremental sales. Looking closely at that model raised deeper questions: what artifact is actually being produced, how is it verified, who funds the reward, and which parts of the system can realistically be decentralized given the constraints of real-world transactions. That analysis ultimately led to identity-free spend tokens and the protocol primitives behind Crinkl—structures that closely resemble the Internet Labor Market framework, not by design but by arriving there through reverse-engineering a working product.
This framework becomes fully visible in what the system calls Wildfire, where brands express demand over verified spend. Receipts are the raw input, but once normalized and verified they become signed spend tokens—artifacts that can be reused as proof surfaces. Brands can define campaign conditions over them, such as category spending thresholds or repeat store visits, and zero-knowledge proofs verify whether those conditions are met without revealing the underlying receipt or the shopper’s identity. The result is a composable market where the unit of work is not merely scanning a receipt but producing a verified economic fact that can generate rewards multiple times. Contributors simultaneously create the dataset and constitute the audience it targets, meaning the participants are not just supplying labor to the system—they are the market itself.
