FitPreview on-device photo preflight
TensorFlow.js Core/Converter/WASM 4.22.0, Copyright Google LLC, Apache-2.0.
MoveNet SinglePose Thunder TFJS version 4, Google, Apache-2.0.
Source: https://github.com/tensorflow/tfjs
Model: https://tfhub.dev/google/tfjs-model/movenet/singlepose/thunder/4
Model card: https://storage.googleapis.com/movenet/MoveNet.SinglePose%20Model%20Card.pdf
Runtime entry: scripts/pose-runtime-entry.js (rolldown, browser ESM).
WASM files copied without modification from the pinned npm package.
Model.json and shards downloaded from the above model URL, /<filename>?tfjs-format=file.
All inference runs locally. The photo and landmarks are not uploaded by this module.
Same-origin CSP restricts requests to this service, not Google/CDN endpoints.
A heuristic preflight, not a guaranteed body classifier, multi-person detector or security boundary.

SHA-256
3697aed96883bd80ebc21f9d467013c6b21fad1368e91f2b83b4be11e82d11f6  runtime.js
74d25aebb524586ab2fa5067cdff444fcbafcf4b44d54df17ed3631467ba43b4  model.json
58dd47fd600a4849c342ce14e2ec32102744ca3d5f7fd9e6888d284f7f922cba  group1-shard1of3.bin
08da980bc00886854f29b49b9c14528cd41626c0498c67b450de394713f1d0bc  group1-shard2of3.bin
c1c6b712e33456aa95aa556b6425528d04fe8730ce9b92de38ec58e947fede34  group1-shard3of3.bin
70a5d516060464e5269f01c74bac1772d6b8ab6cb612acf16b5cdaf61f78d892  tfjs-backend-wasm.wasm
77ebb28a6d34f371dbbf2086b7f2de8994acd8ea5a3cf1fa24d2c26c840cac7b  tfjs-backend-wasm-simd.wasm