Security boundary
Live WeChat data is treated as read-only input. Sensitive decrypted copies, keys, exports, caches, and local app state are excluded from publication and kept outside the public site.
A local, read-only macOS workbench for deciphering WeChat 4.x database copies, querying message/session structure, and building privacy-bounded analytics. This page documents architecture only; it does not expose real messages, keys, private paths, screenshots, or account data.
Live WeChat data is treated as read-only input. Sensitive decrypted copies, keys, exports, caches, and local app state are excluded from publication and kept outside the public site.
The workflow separates identity inventory, key extraction, read-copy decryption, local text API probing, export tooling, and dashboard/app layers so each stage can be verified without writing to the live data directory.
The analytics path standardizes message fields, session manifests, voice transcript schema, and dashboard outputs while preserving a strict no-commit boundary for private material.
Private-data engineering discipline: local-first tooling, read-only boundaries, staged pipelines, schema documentation, desktop packaging, and analytics design without exposing the underlying personal corpus.