In monitoring mode the bot logs every action it would take but does not actually delete, mute or ban. It is the safest way to trial Telm on a live group: watch the journal on your real traffic, confirm the decisions match your judgment, tune anything that looks off, then turn monitoring off to enforce for real. Nothing in your chat changes while it is on.
Want to set this up? Do it right in your Telm account.
Open in your dashboard1What monitoring mode is
Monitoring mode (also called silent, observe or dry-run mode) is a safe way to try Telm on a real group. The full detection engine runs on every single message — spam scoring, CAS and global blocklist checks, AI moderation, your rules and triggers — but instead of deleting, muting or banning, the bot only records what it would have done.
This lets you see how the bot behaves on your actual traffic, with your real members and your real spam, before it touches a single message. Think of it as a test drive with the engine running but the brakes and steering not yet connected to the wheels.
- Every message is still analysed exactly as it would be in enforcement.
- No message is deleted, no member is muted, kicked or banned.
- Each decision is written to the moderation journal with the action that would have been taken.
2How monitoring works
When monitoring is on, Telm reaches a verdict for each message — allow, delete, mute, kick or ban — together with a reason and a confidence score. The only difference is the final step: enforcement is skipped and the verdict is logged instead of applied.
Because the same engine and the same settings produce the log, what you see in monitoring is exactly what you will get in enforcement. Turning monitoring off simply reconnects that final step — nothing about how decisions are made changes.
- Detection, scoring and rule matching are identical to live mode.
- The verdict, the reason and the confidence are all recorded.
- Switching to enforcement changes nothing about how decisions are made — only whether they are carried out.
3Turning it on and off
Monitoring mode is a single switch in the group's settings. While it is on, nothing is enforced — the group runs exactly as before, but the journal fills with the decisions the bot is making.
When you are confident, turn monitoring mode off. From that moment the bot enforces its decisions for real, using the very configuration you just validated. You can switch back to monitoring at any time if you want to test a big change before it goes live.
4Why and when to use it
The point of monitoring mode is trust. A moderation bot that can delete messages and remove members is powerful, and the fastest way to be comfortable handing it that power is to watch it work with no risk first. A day or two of monitoring on a busy group usually shows enough decisions to know it behaves the way you expect.
It is useful far beyond the first day, too. Any time you make a meaningful change — apply a new preset, add strict custom rules, raise the strictness level — monitoring lets you see the effect on real traffic before it can remove a legitimate message.
- Onboarding a new group and building confidence before going live.
- Testing a new preset or a batch of custom rules on real messages.
- Checking a stricter configuration in a sensitive or high-value community.
- Reassuring co-admins and members that the bot will not over-moderate.
5Reviewing the results
Open the moderation journal to see every decision: which messages were flagged, why, what action the bot would have taken, and how confident it was. Filter by action or reason to focus on the deletes and bans first, since those are the decisions that would have had the biggest impact.
If you spot a false positive — a legitimate message the bot would have removed — adjust the relevant filter, rule or whitelist before going live. If everything looks right, disabling monitoring mode enables enforcement with the same configuration you just validated.
- Start with the would-delete and would-ban entries — the highest-impact actions.
- Open a decision to see the exact reason and which check triggered it.
- Fix false positives with whitelists, trust settings or a lighter strictness level before enforcing.
6Best practices
Monitoring mode is most useful when you actually review the log and act on it. Leaving it on for weeks without looking means the group is unprotected the whole time — real spam is being logged, not removed.
Give it enough traffic to be meaningful, then commit. A quiet group might need a few days to accumulate decisions worth judging; a busy one can be validated in hours.
- Set a clear finish line: review the journal, fix what is wrong, then switch to enforcement.
- Re-enter monitoring for a short spell after any big configuration change, not permanently.
- Pair it with a protection preset so you are validating a sensible starting point, not an empty configuration.
7Common mistakes to avoid
The single biggest mistake is treating monitoring mode as a finished setup. It protects nobody — while it is on, spam that the bot flags is still sitting in your chat. It is a rehearsal, not a live show.
The second is judging it by silence. If the journal is empty, that usually means your group simply has not received spam yet, not that the bot is broken. Test detection on real traffic over time rather than expecting an immediate flurry of decisions.
- Do not leave monitoring on indefinitely — nothing is being enforced.
- Do not conclude the bot is inactive just because the journal is quiet on a calm day.
- Do not skip the review step — the whole value is in reading the log and adjusting.