Questions, answered.

Yes. The Layer-1 classical engine is CPU-only and costs effectively $0 per message to run, so M1 has a genuinely free tier — not a trial. Premium and Enterprise tiers add higher limits for larger communities, and the paid M2 tier adds LLM-grade detection for AI-driven attacks.

You'll know long before it happens: admins get a heads-up at 80% and repeated reminders from 95%. On the free plan, protection pauses at the limit until you upgrade or the month resets — your group sees a single status line. On paid plans, burst credits kick in: screening keeps running at full strength past your quota, and you set the ceiling yourself — a message limit or a time window ("protect me 8 hours into a burst, whatever the volume"). You're alerted the moment a burst starts, and protection only pauses at the point you chose. Raise the limit or upgrade and it resumes within a minute.

Request access at app.targa.dev/register (private beta), then add @HeyTargaBot to your group as an admin and connect the channel from your dashboard.

Two: Delete Messages and Ban Users — that's what removing spam and restricting spammers require. Everything else can stay off: Change Group Info, Invite Users via Link, Pin Messages, Manage Topics, Manage Video Chats. Leave Add New Admins and Remain Anonymous off — @HeyTargaBot never needs to escalate its own rights, and a security bot should be visible. Set its member tag to “bot”: promoting any bot flips Telegram’s corner label from “bot” to “admin”, and a security bot shouldn’t wear the label impersonators imitate. You can also add it with no admin rights at all: it stays a passive observer and still computes your Targa Score at member depth.

Each audience gets its own command menu: members see only what they can use — /report to summon admins discreetly (the report is hidden from other participants, so no one sees who reported), /rules, and /why to read the reason behind any moderation action. Admins get the full menu, led by /score. No member ever sees moderation commands they can't run.

Every action carries a structured reason. Send /why with the verdict id and you get the verdict, the score, and the exact signals that fired — for example avatar_pHash_matches_admin · join_age<10m · first_message_link.

Homoglyph-obfuscated text, edit-after-post bypasses, admin impersonation, link/crypto giveaway spam, and flood bursts. The M2 tier extends detection to conversational scam funnels and hyper-personalized phishing.

No. Targa focuses on screening messages and members; rule-based moderation bots can run alongside it.

No — closed source, and hosted by us. You don't run infrastructure, and detection internals stay ahead of adversaries.

Messages in groups it protects, as any admin bot does. Verdicts and signals are retained so /why works; see the dashboard's privacy page for specifics.