BotGuard — Conversation Check

Local comment-pattern checks on supported sites. A flag is not proof of identity.

BotGuard provides experimental local checks of comment patterns on public posts. Enable monitoring with your consent and permission for each supported site. It then checks comments already loaded on the page and shows inline status labels and a summary in the extension popup. A manual check is also available. This beta checks public comment views on supported sites that you enable in Settings. Only loaded, safely attributable comments are assessed; some layouts, short comments and media-only comments may not be assessed. It does not automatically scroll, post, message or block accounts. Private messaging routes are excluded. With your consent, BotGuard processes comment text, names, account identifiers, timestamps, links and page context locally to identify repeated wording, timing patterns and related indicators. Reports, identifiers, derived fingerprints and feedback may be stored in your browser profile. New comment text is not retained. No scan data is sent to a shared server, external AI or translation service. You can stop monitoring, manage site permissions, export or delete local data in Settings. Exported backups are not encrypted. See the privacy policy for retention and storage details. The interface supports Hebrew, English, Spanish, French and Arabic. It follows the browser language with English fallback, and you can select a language in Settings. Original comments and names are not translated. A flag or percentage is an experimental rule-based indication for review, not a probability or proof that an account is a bot, uses AI, is fake or is fraudulent. No flag does not verify a human. Detection accuracy on current supported platforms has not been validated. Version 9.7.5 expands public-comment support and improves automatic updates, attribution and status-label display. BotGuard is independent of the supported platforms and is not endorsed by them.