The moderation journal is the full record of what the bot did and why. Every automatic action and every manual admin action lands in one place, tagged with its category, score and the layer that decided — filter, trigger, custom rule, language policy or the anti-spam engine. You can search, view statistics and export to CSV. Labelling detections as spam or false positive trains your group's protection, and Monitoring Mode lets you watch every decision before you enforce.
Want to set this up? Do it right in your Telm account.
Open in your dashboard1What the journal is
The moderation journal is your single source of truth: who was moderated, when, for what, and by which layer of the system. Automatic bot actions and manual admin actions live in the same log, each tagged with its source, so there is never a mystery about what happened.
It is where trust in the bot is built. Nothing happens silently — every deletion, warning, mute and ban is recorded, can be reviewed, and can be undone. If a member ever asks why their message disappeared, the answer is one search away.
2What gets recorded
Each entry captures the message and the decision behind it:
- Category — why it was flagged (scam, promo, foreign language, a stop-word, and so on).
- Score — how strongly the message was judged as spam.
- Layer — which part of the system acted (a filter, a trigger, a custom rule, the language policy or the anti-spam engine).
- Source — whether it was an automatic action or a manual one by a named admin.
- Action — what was done: delete, warn, mute, ban or kick.
3Reading the layer that decided
The layer tag is the most useful field when you are tuning. It tells you whether a removal came from your own configuration or from the shared engine, which points you straight at the setting to adjust.
If a catch is tagged to a filter, a trigger or a custom rule, the fix is in your own settings — loosen the filter, retune the trigger, tighten the rule. If it is tagged to the anti-spam engine, the right move is to label it (see below) so the system learns, rather than changing a setting.
4Review and label detections
The journal is not just a log — it is how you teach the system. Open a detection and mark it: confirm spam on a genuine catch, or mark as false positive on a mistake.
Every label trains the protection specifically for your group, so the bot gets steadily better at your community's real traffic. Confirmations reinforce good catches; false-positive marks pull the system back from patterns that were hitting real members.
5Search, statistics and export
The journal supports search across entries and shows statistics on what is firing most — which categories dominate, whether false positives are creeping in, how busy moderation has been. Use it to spot a noisy rule or a filter that is over-reaching.
You can export the full log to CSV for your own records, reporting or an external audit. Everything you see in the dashboard can leave it in a spreadsheet-friendly form.
6The safe rollout workflow
The journal is the centre of the recommended rollout for any new group. First, keep Monitoring Mode on: the bot logs what it would remove without actually acting, so you can watch it work with zero risk to real members.
Then review the journal, label the detections, and only switch to real enforcement once you are confident. Even after enforcement, every action stays in the journal and can be undone, so the safety net never goes away.
7Using the journal to troubleshoot
The journal answers the two most common questions owners have. When a member asks why was my message removed, search their name or the text and read the category and layer — see why was a message removed. When spam is getting through, check whether it appears in the journal at all: if it does with a low score, label it; if it does not, the bot may lack the rights to act — see bot not removing spam.