A quantum circuit.
Simulated in your browser, byte-for-byte.
What arrived over the wire is a wai.quantum.circuit container — a
gate op-log, a few hundred bytes. Your device classically simulates it with the
same pure-i64 Rust native sinks run, compiled to wasm32: every amplitude is a
fixed-point complex integer (re, im : i64), no float anywhere, no ML
runtime, and no quantum hardware. The circuit is tiny; the state it reconstructs is
2n amplitudes — exponential. Because the whole simulation is integer, the
reconstructed statevector is byte-identical on every machine, so the page re-checks its
BLAKE3 against the hash the native reference produced (statevector-equivalence). The
two QFT presets are the quantum Fourier transform — the “quantum FFT,” carried
as a circuit and reconstructed exactly. No speedup is claimed: this makes a quantum
computation a transportable, receiptable object, not a fast one.
Why this is worth a receipt
Real quantum hardware is noisy and its measurement is stochastic — the opposite of the
byte-exact determinism WAI rests on. So WAI does not run on a quantum computer; it
carries a quantum computation. The op-log above is a few hundred bytes; your device
reconstructed the exponential statevector and proved its hash matches the native reference
and the conformance corpus — the same bytes on every machine. Statevector simulation is
exponentially expensive, which is exactly what makes an energy-accounted, byte-exact,
signed record of it a genuinely new artifact (see the QuantumReceipt). No
quantum advantage is claimed — the value is that the result and its cost are auditable.
Reference engine + conformance corpus in the open-standards repo, Apache-2.0.