Content Moderation That Never Sees Your Private Messages

Описание и применение

SifEdge обеспечивает структурную защиту приватных сообщений, полностью исключая их из классификатора и серверных логов.

Задача

Moderation systems often claim to respect private messages, but in practice the content is still processed by a classifier before any privacy boundary is applied, exposing private data to third‑party services.

Результат

Private messages are never processed by a classifier or sent to a backend, providing a structural privacy guarantee that cannot be bypassed by misconfiguration.

Рабочий процесс

• 1. Integrator sets channelContext to 'private' for private channels.
• 2. SDK’s checkText function checks channelContext; if 'private', it returns ALLOWED without invoking the classifier.
• 3. If channelContext is not 'private', the classifier runs locally via transformers.js.
• 4. Server receives decision payload; if channelContext is 'private', it returns a 400 error and refuses to log or store the result.

Доказательства

The gate is the first line, not a filter on the last line. There is no code path in the SDK where a private message reaches a classifier, because the function returns before the classifier is even referenced.

Источник

If a developer misuses the client SDK, or wires up a custom integration that skips it, the server still refuses to accept the result.

Источник