silentpayments.net
Detail
plain words, no jargon the standards, named every number, and how to check it

§ 00 · The north star

Kill the bitcoin address.

The 42-character single-use address is a fifty-year usability failure. Bitcoin can do what every payment system before it did: one identity for life, private by default, with nobody to trust, eventually.

This page is that destination, the fixed point the roadmap and future specs point back to.

§ 01 · The problem, honestly stated

Receiving privately is the hard part.

Silent payments (BIP 352) give you a reusable address; finding your money means scanning everything, so a helper does the heavy lifting.

The weak link

A helper you must trust can go silently wrong: one major wallet’s backend did in 2024. The worst designs learn everything about you.

Physics first: that data needs full-node data, so “no helper” can’t be literal. It means no one you must trust, then no one in particular.

§ 02 · The ladder

Four stages, each one shrinking the trust

1 · A server whose lying is provable, running now

Every answer is fingerprinted into a public chain, checkpointed beyond the operator’s control. Cheating becomes impossible: this is the lookout, live, proven with real money.

2 · Many servers, one standard, in progress

One protocol, matching fingerprint heads: servers become interchangeable, checking each other so none needs trusting. The converged protocol draft is public, built on measurements nobody else has.

3 · The node network serves it, the class dissolver

Nodes serve scanning data over the peer-to-peer network, same as block filters today (BIP 157/158; a Core tweak index exists as an unmerged branch). The helper class dissolves; fingerprints double as lying-peer detection, the same structure our chain already has.

4 · The destination, committed by consensus

A soft fork commits each block’s scanning fingerprint into the chain itself (BIP 158’s reserved-hook precedent: commit only what’s specified and deployed). The data cannot be wrong, whoever hands it to you; servers become couriers. Nobody, including us, has proposed this yet: it needs stages 1 to 3 first.

No rung skips the others: consensus only commits to what’s already specified, measured, and running. The boring work is the path.

§ 03 · The honest edges

What even the endgame does not fix

Consensus commitment makes served data impossible to fake. It does not promise magic. Three edges stay open at every stage.

Behaviour is still visible

Real dust sent to you, and watching who fetches what, both use honest data. Wallets keep their own defences; the ladder only fixes receiving trust.

Notifications are a shortcut, not a rail

Sender notifications can make scanning the fallback, not the daily path. Never a replacement: the senders that matter most, a donation page or an exchange, won’t cooperate.

One lane stays closed

A server that finds your payments without learning they are yours needs a sender tag we don’t have, or ~100 machines per user in curve math (OMR, FMD, PIR, homomorphic encryption). The arithmetic. The phone scanning itself is the ceiling, not a compromise. A blind push (one hourly message naming new blocks to every phone) closes the gap: the server learns nothing.

§ 04 · Why this kills the address

Names finish what the ladder starts

A reusable address, safe for anyone to receive on and trusting no one, beats a string you must rotate. Put a human-readable name on top (BIP 353, both rails) and the address becomes what the IP address became: still underneath, never typed. That is the whole bet of this site.

§ 05 · The receipts

Where each claim lives

Written 2 September 2026, the day stage 1 opened to the public over Tor. Advances get recorded in the change log.