TG AI Bot

Reply Settings

Content Filter

Control how the bot handles spam, adult content, and prohibited messages.

Filter Mode Off: All messages saved.
Filter: Bad messages not saved.
Delete: Bad messages removed from chat.
Delete message and block user: Delete the message and ban the sender (bot needs admin with Delete messages and Ban users). Use AI Filter AI catches sophisticated spam (uses LLM) Trusted sender min messages (0 = disabled) Sender with this many messages in the group is "trusted". New authors get full check; trusted get light or no check (spam is often from new accounts). Filter for trusted senders Full: AI + regex. Light: blocked words only. Off: allow without checking. Trusted author IDs (Telegram user IDs) These users are always treated as trusted even with 0 messages (e.g. moderators).

Note: For Delete the bot must be group admin with Delete messages. For Delete message and block user it also needs Ban users.

Quick Actions

Qdrant indexing is paused while Reply Engaged is off for this group. Messages will still be stored and spam will be filtered/deleted, but they will not be indexed for retrieval. After enabling replies, run Reindex Qdrant to rebuild the search index from DB.

View Stats

Reindex Qdrant: Rebuilds the search index from DB using content-based IDs (one vector per unique content; repeated ads merged). Use after enabling replies (since live indexing is paused when replies are off), and after large sliding-window/content-dedup changes.

Group Settings

Папы🇺🇦 Варны 🇧🇬 Общение

Reply Settings

Content Filter

Quick Actions