Silent Payments (BIP352)

A receiver publishes one static address (silent payment address). Senders compute a unique fresh output for each payment from the address + their own keys via Schnorr-based ECDH-like derivation.

No notification tx. No address rotation. Strong privacy by default.

Mechanism

Receiver

Has scan key b_scan and spend key b_spend. Publishes:

silent_address = encode(B_scan, B_spend)   bech32m
                  = scan pubkey + spend pubkey

Sender

For each payment:

Generates ephemeral private a (or uses an input's privkey if available + safe).
Computes shared secret with receiver: S = ecdh(a, B_scan) = a * B_scan = b_scan * (a*G)
Tweaks receiver's spend key: P = B_spend + Hash(S || k) * G for k = 0, 1, 2, ...
Sends to P2TR(P).

Receiver finds outputs via:

For each tx in chain, compute S = b_scan * input_pubkey.
Check if any output matches B_spend + Hash(S || 0) * G.
If yes, derive private key for the output.

Privacy properties

No public address per payment — analyst sees random P2TR outputs, not a static address.
No notification tx (vs BIP47).
No correlation between different senders to same recipient.
Forward security — past outputs untraceable from future payments.

Scan cost

The trade-off: receiver must scan every Taproot output of every tx for matches.

For a node already running: marginal cost.
For a light client: needs a server to source the per-tx tweak data; BIP158 filters alone do not help, because the candidate output keys are not known until the tweak is in hand. BIP352 Appendix A ("Light Client Support") leaves this explicitly as open research.

Two lines of work as of September 2026. The BIP0352 Index Server Specification (silent-payments/BIP0352-index-server-specification, still WIP, last pushed May 2026) standardises three server stacks — Remote Scanner (ephemeral), Tweak Server (anonymous), My Scanner (personalised) — over the already-deployed blindbit-oracle, cake-esplora and Frigate indexers. Separately, BlindBit Oracle v2 drops filters entirely and streams per-output data; Rob Segers' benchmark against taproot-only filters is in Optech #422 (2026-09-11).

Implementations

Status as of September 2026.

Sparrow — send to sp1 addresses since 2.3.0 (October 2025); BIP375 PSBT fields plus DLEQ proof verification for hardware-wallet sends in 2.4.0 (February 2026); full receiving wallets, incl. airgapped signers, in 2.5.0 (May 2026).
Frigate — Sparrow's Silent Payments Electrum server. Scans server-side with ephemeral in-RAM client keys, per the "Remote Scanner" model of the BIP352 Index Server Specification; optional GPU backends. Sparrow 2.5.0 ships frigate.2140.dev as a public SP-capable server.
Cake Wallet — on-device receiving and scanning, with scan-from-height / scan-from-date controls and an automatic switch to an SP-capable node while scanning.
Wasabi — send only; SP addresses resolve in the send path, no scanning.
Nunchuk — sending, via silentpayment.hpp in libnunchuk.
Electrum — no upstream support; SP arrives via third-party plugins (e.g. electrum-silent-payments-sender), not a core feature.
libsecp256k1 — silentpayments module, added upstream in v0.8.0 (August 2026). Sending and receiving; requires full transaction data (no light-client scanning).
bdk_sp — experimental BIP352 crate (0.2.0, July 2026). The silent-payments-* crate family (core / send / receive / scan / psbt / descriptor) is a separate, also pre-1.0, Rust stack.

Bitcoin Core has not shipped wallet-level silent payments. The original PRs (#28122 "Implement BIP352", #28201 "sending") were closed unmerged on 2026-05-12 and superseded by #35301 "Implement bip352 (take 2)" and #35302 "Sending (take 2)", both opened 2026-05-16; #35302 and the older #32966 "Receiving" (opened 2025-07-14) are still drafts. Core master carries the libsecp256k1 silentpayments module via the 2026-07-23 subtree update, but no release up to and including v31.1 (July 2026) contains it.

Two supporting BIPs are still Draft as of September 2026: BIP375 (sending silent payments with PSBTs) and BIP392 (sp() output script descriptors). BIP352 itself is status Complete.

Hardware wallets: no longer a blocker for sending — BIP375 carries the SP data in the PSBT and the signer proves its ECDH share with a DLEQ proof (shipped in Sparrow 2.4.0). Receiving still needs b_scan on a host that can scan; an airgapped signer holds only b_spend.

Compared to BIP47

Aspect	BIP47	Silent Payments
Notification tx	Yes (one per pair)	No
Per-tx scan cost	Low (only your derivation)	High (every output)
Static address	Per-pair	Universal
Privacy	Good	Better
Standard	BIP47 (assigned 2015), Deployed	BIP352 (assigned 2023), Complete
Wallet adoption	Sparrow, Stack	Sparrow, Cake; Wasabi/Nunchuk send

Adoption row as of September 2026.

Use cases

Charity / donation addresses publicly posted.
Newsletter sponsorship receivers.
Recurring deposits from many parties.

Limitations

Scan cost — not trivially Lightning-light-client friendly. In practice wallets either lean on a scan/index server (Frigate) or accept long on-device scans (Cake Wallet).
Hardware wallets — sending is solved (BIP375 PSBT fields + DLEQ proofs, Sparrow 2.4.0); receiving still needs the scan key online.
No Bitcoin Core wallet support — as of September 2026 the BIP352 wallet PRs are still unmerged; see Implementations.
New — ecosystem still maturing.

bitcoin-privacy-silent-payments