A privacy design moves into a public draft

Konclave disclosed an experimental Oblivious HTTP prototype on November 15, 2026, giving developers a public draft to inspect as it works on separating users' internet addresses from requests to its own servers. The project's pull request explicitly says the experiment is undeployed and is not part of the application.

The development surfaced in a Zcash Community Forum discussion about wallet infrastructure. A project contributor linked the new draft and identified streamed responses as unfinished work. That makes this a concrete development artifact, rather than a general promise of better privacy, while leaving a substantial gap between an experiment and a feature people can rely on.

Why a relay can change what a server learns

Oblivious HTTP, or OHTTP, is an Internet Engineering Task Force standard for forwarding encrypted web requests. The basic arrangement separates the party that receives the client's connection from the party that can read the protected request. A forwarding service carries a sealed message onward instead of learning everything from one direct connection.

The distinction matters because privacy has several layers. Protecting transaction information does not automatically remove the information exposed when a browser contacts a service. An internet address and the timing of requests can raise different questions from the contents of a blockchain transaction. Readers exploring Zcash and its privacy model should keep those layers separate.

This is also why an OHTTP label is insufficient on its own. The relevant questions include who operates each stage, which requests take that route and what happens when a connection fails. A diagram showing two servers cannot establish that their operators are independent.

The experiment has a deliberately narrow boundary

According to the draft, the tests involve reads from Konclave's coordinator, with no blockchain transactions, wallet operations or funds involved. The project has not selected an independent relay operator. Its description also flags direct requests and the combination of operators' records as limits to the proposed separation.

The repository describes Konclave as a browser-based collective treasury project using threshold signatures on Zcash. It also warns that the product has not been independently audited and identifies continuing trust in its hosted coordinator. Network-privacy work therefore addresses one part of a larger design, not every risk associated with collective custody.

That distinction helps explain what a successful test would mean. Showing that a sealed request can reach its destination is useful engineering evidence. It does not establish that an entire application's traffic follows the same path or that its custody arrangements have been independently assessed.

What would make the next milestone meaningful

The next useful evidence would connect the experiment to a defined operational deployment: a clearly identified operator arrangement, a description of covered traffic and reproducible checks of the remaining exposure. A later release should be assessed on those terms, rather than on the existence of this draft alone.

For readers comparing crypto infrastructure, this is the same discipline needed when evaluating the separate components of a cross-chain transfer: identify the particular problem being addressed, then examine the assumptions left elsewhere. Konclave's public prototype gives that discussion a specific starting point. It does not yet change the privacy guarantees of the live application.