04 Security · Q-Entangle-Bell
Fold a witness into the key. Never a floor beneath it.
After both peers complete the Entanglement Ceremony, a session can fold a CHSH-verified, Tsirelson-window-validated Bell-game transcript both peers seed and any auditor can verify into its key derivation — off by default, above the hybrid floor.
01The claim — layered above the hybrid floor
Bound above the floor. Never instead of it.
A session key can be bound to a CHSH-verified, Tsirelson-window-validated Bell-game transcript both peers seed and any auditor can verify. The binding is conditional, not automatic: both peers complete the Entanglement Ceremony in the chat, and only then does the witness fold into the Q-Ratchet chain-key derivation — every eight advances, on both sides or neither. Without the ceremony the session keys on the X25519 + ML-KEM-1024 hybrid floor alone, and the code takes that branch by default. Breaking a session requires breaking both X25519 and ML-KEM-1024.
CNSA 2.0 disallows QKD for National Security Systems — a position we agree with. Q-Entangle-Bell is not QKD: it adds a verifiable transcript to the key derivation, and its removal would leave the hybrid handshake exactly as strong.
02Run it — SHAKE256 rounds, SHA3-256 chain
Run the Bell game yourself.
Real math, in this browser: every round re-derives from the two peer seeds via SHAKE256, rounds chain by SHA3-256, and VERIFY CHAIN recomputes the whole transcript. Push the S value past the classical bound of 2 and watch the verdict change.
Run the Bell test yourself.
Two peers seed a CHSH ritual; every round re-derives from those seeds and chains by SHA3-256. Verify the chain, then fetch a signed pulse from the live beacon and verify that too.
PEER A — · PEER B — · FUSED SHA3-256 —
S = —
awaiting rounds
- E(a,b)
- —
- E(a,b′)
- —
- E(a′,b)
- —
- E(a′,b′)
- —
- PAIRS
- 0
a 0° · a′ 45° · b 22.5° · b′ 67.5° — CHSH-optimal settings, one round = 512 pairs, settings switched per pair.
Honesty note: your browser cannot hold entangled photons — this bench samples the exact quantum statistics (E = cos 2(a−b)) versus a genuine local hidden-variable model (λ uniform, deterministic outcomes). The point it proves is mathematical and real: no local model crosses 2; quantum statistics do. Every signed pulse below carries a backend label — today’s beacon pulses say local-sim, and the verifier shows it. When hardware pulses ship, that label is how you will know.
Recompute a signed beacon pulse in this browser
Your browser fetches one pulse, recomputes its SHA3-256 self-commitment from the raw bytes, checks the hash-chain link, and re-derives the CHSH Bell-violation flag exactly as the signer did.
This button is the only network request this instrument ever makes — one GET to https://quantumentanglement.io/pulse/latest, a QERYX-owned origin, and only when you press it.
Press FETCH & VERIFY to pull the latest signed pulse from the beacon and recompute it here.
Until you run the downloadable verifier, a passing widget means self-consistent and chain-linked — not yet signature-authenticated. The bytes are identical; only the lattice-signature math is deferred.
03The two lanes — v1 in research, v2 designed
Tell the two lanes apart. We name both.
v1 — in research: the deterministic dual-peer CHSH ritual
Both peers seed a deterministic CHSH ritual computed by SHAKE256 simulation. It is a simulation and the protocol labels it one: its value is that the 623-byte transcript is a deterministic function of the two seeds, re-derivable by anyone holding them, and unchosen by either peer alone — not that qubits were measured.
v2 — designed, not active on any deployment: operator-signed QPU transcripts
The v2 lane would bind keys to operator-signed transcripts from real quantum processors. The construction, the acceptance window and the operator-signature verifier are written in crypto-core. No deployment serves an operator transcript yet, so the lane is designed and not active; the v1 ritual carries the binding.
Downgrade protection
Every downgrade records its reason — peer contribution absent, CHSH below the floor, CHSH above the ceiling — and the relay publishes the reason as a transparency counter. Each entropy refill records the name of every source that contributed and the reason any source was rejected.
The deeper physics — entanglement, the CHSH game, why 2√2 — lives on /science and the dedicated instrument site, quantumentanglement.io.
04The record — the questions we get asked hardest
The hard questions.
Why add Bell-binding at all if the hybrid floor already holds?
Because a key-derivation input that any auditor can re-verify, and that neither peer can choose alone, is a different kind of evidence than a hardness assumption. It costs 623 bytes per session and removes nothing: the floor beneath it is unchanged.
Does a simulation add real security?
The v1 ritual adds verifiability, not entropy hardness — the honest claim is auditability of the transcript, and that is the claim we make. Key entropy comes from the hybrid handshake; the conditioned pool is folded into group sender-key roots and does not enter a one-to-one message key. The v2 QPU lane is where physical measurement would enter, and we name it as a separate lane precisely so the two claims never blur. That lane is not active on any deployment.