An open protocol for AI-mediated product discovery

Niklas Henkel released Draft v1.0 of the Open Receipt Protocol (ORP), a public-good protocol for accountable AI-mediated product discovery, on 2026-08-07. The protocol uses signed rating receipts tied to specific product claims, verified human judgments, and a neutral non-profit operator to build public graphs of supply, demand, and accountability while keeping conversations and identities private. The white paper and technical overview are available under CC BY 4.0, and adversarial review is invited via GitHub issues.

Competing on Quality, Not Attention : accountable product discovery through AI hosts. ORP Open Receipt Protocol is a reciprocal public-good protocol for accountable discovery. Companies register claims about their products. LLM Hosts facilitate recommendations and feedback. Verified humans contribute the claim-specific judgments. Those contributions build graphs of supply, demand and accountability. The structure, commitments, and aggregate signals of all three are public; individual conversations and identities stay private throughout. A neutral non-profit operator or a federation of such maintain the graphs but may sell no influence over them. The contribution is the accountability loop: signed rating receipts that reference the exact claims a recommendation was made on, produced by verified humans, tied to a recommendation the operator itself served, and held by an operator with nothing to sell. A host that runs its own version of this cannot write these receipts, because everyone can see whose thumb is on the scale. This is what makes the quality signal hard to fake and hard to buy retroactively. One-page abstract /open-receipt-protocol/orp/blob/main/ABSTRACT.md PDF /open-receipt-protocol/orp/blob/main/abstract.pdf : start here. White paper /open-receipt-protocol/orp/blob/main/whitepaper.md PDF /open-receipt-protocol/orp/blob/main/whitepaper.pdf : the full thesis. The centralization problem, the parties and their incentives, the core mechanism, identity and privacy, scoring, the transparency model, operator neutrality, and a per-surface attack map.: the engineering-depth companion. How each concept works, at the depth needed to build against and to attack. Technical Overview /open-receipt-protocol/orp/blob/main/technical-overview.md Draft v1.0 thesis 2026-08-07 . In the paper's own words: it is "a draft contribution written in a finished register, not a finished architecture: it argues a thesis and shows one way to build it, and it does not pretend to have found every flaw in its own design or an optimal solution. It invites the scrutiny a claim-accountability protocol should hold itself to." Adversarial review is welcome: open an issue https://github.com/open-receipt-protocol/orp/issues . Niklas Henkel. Contact via issues https://github.com/open-receipt-protocol/orp/issues . Text and figures are licensed under CC BY 4.0 /open-receipt-protocol/orp/blob/main/LICENSE .