NISHI SHORE - REVIEW STAGE b1787917928

the flat-light REVIEW STAGE: mixed beachwear across the cast (bred, stable per girl) · same engine, same controls as NISHI CRAFT · ?review=1 auto-frames the nearest girl
ISEKAI: cross to NISHI CRAFT or NISHI VALE — level, ledger and party come along
TISSUE PHYSICS, DERIVED NOT TUNED — every figure here is COMPUTED by the shipping solver as this page is written, so it cannot drift from the code running above it. Chest plant 4000–5200 mHz across bred bodies (cited tissue_fn_mhz 4000–5200, free-vibration modes) · damping 480 permil (haake-scurr-2010 measured 475, and what our own XPBD ringdown read) · downward stiffness 89 tenths of the upward branch (cai2018 piecewise, selected by one sign test) · anisotropy AP 610 / ML 600 permil of vertical (mills2025) · 1001 distinct bodies drawn from each girl's whole genome, where there were 4 · proven by nx_gamefeel_oracle_gate and nx_softbind_gate.
Every pixel above is emitted by NishiLang compiled to WebAssembly.
This page holds no game logic — every rule, every position and the quality decision itself live in NishiLang compiled to WebAssembly. The page picks the best renderer your device actually has: the GPU when one is available, and the sovereign NishiLang rasteriser otherwise, which is also what you get with ?gl=0. Both draw the same organ state.
See it yourself: sovereign render — the NishiLang-drawn world and characters · GPU render — device resolution · WebGPU render — experimental, world only for now (the character pass on WebGPU is a named owed rung; the default WebGL2 tier renders the full cast)
Visiting characters — the estate's reference rigs with authored surfaces carried in-asset (TEXC+TEXM, gate-proven) — are OPT-IN while their render is repaired: add ?cast=1 to load them. They are not on the default page because the arrival is slow and the figures do not yet render correctly — a visibly broken feature on the front door is worse than an absent one.