Telm ships three ready-made strictness presets — Light, Medium and Strict. All three keep the core protection on (global spammer database, AI moderation and newcomer scrutiny); they differ in how many mechanical restrictions sit on top. Light stays out of the way of a lively chat, Medium closes the common spam vectors without friction for honest members, and Strict turns the group into a fortress with captcha, a 24-hour newcomer quarantine and links allowed by whitelist only. Pick a level in seconds, then fine-tune any individual switch.
1One dial, three positions
A strictness level is a bundle of settings applied together, so you do not have to reason about dozens of switches one by one. Pick the position that matches your community and Telm configures the rest.
Whatever you choose, the same foundation is always on: the global spammer database (CAS) blocks known spammers at the door, AI moderation reads meaning, and newcomers get extra scrutiny on their first messages. The levels only change how strict the mechanical rules on top are.
- Light — the softest option: core protection only, minimal friction.
- Medium — the recommended default for most groups.
- Strict — maximum protection for high-value or heavily targeted groups.
2Light — keep it conversational
Light leans almost entirely on the intelligent layers and adds no mechanical restrictions. It is the best choice when false positives would hurt more than the odd spam message — support chats, small trusted communities, freewheeling discussion groups.
- Global spammer database, AI moderation and newcomer guard on.
- No captcha, no newcomer restrictions.
- Forwards, links, external promotion and crypto deals all allowed.
- A higher warning threshold before anyone is punished.
3Medium — the recommended default
Medium closes the spam vectors that hit almost every public group, while staying invisible to honest members. This is the level we recommend for most communities.
- Blocks invitations to move into private messages (a classic spam pattern).
- Blocks promotion of third-party channels and bots.
- Blocks P2P and OTC crypto deal adverts.
- Catches repeated (duplicate) spam messages.
- Turns on the anti-raid shield with a captcha challenge and notifies the owner on a join surge.
- A moderate warning threshold before punishment.
4Strict — a fortress
Strict layers every mechanical control on top of the intelligent engine. It suits groups under constant attack or where a single spam message is unacceptable — paid communities, launch groups, high-value crypto chats.
- Captcha required on join.
- Newcomers quarantined for 24 hours: no media, no link previews, no inviting others.
- Links allowed only from your whitelist; forwards, channel posts, inline bots and stories blocked.
- Rate limiting against flooding.
- Anti-raid shield with full lockdown on a join surge.
- The lowest warning threshold before punishment.
5Fine-tuning after you pick a level
A level is a starting point, not a cage. After choosing one you can flip any individual switch — allow forwards on Strict, add a captcha to Light, whitelist a partner channel — and your changes stick on top of the preset.
The onboarding wizard can also suggest a level from a plain-language description of your group, then apply only the specific rules you explicitly asked for.