| Tool | Best at | Weak for | Pricing model |
|---|---|---|---|
| Rose | Keyword rules, captcha, anti-flood, custom commands | Semantic scam understanding; doc-based FAQ | Freemium + paid tier |
| Combot | Engagement stats, leaderboards, warnings, web dashboard | Novel scam wording; multilingual support | Freemium + paid tiers |
| Safeguard | Anti-raid verification gates, join bursts | Not a scam-content reader | Freemium |
| GroupButler (AI-native) | Semantic scam interception with reason codes, FAQ from your docs, 14+ languages, on-chain checks, public Arena evidence | No leaderboards/gamification (pair with Combot if you need them) | By total members: Free ≤200 · $9.9 ≤1K · $29 ≤5K · $99 ≤20K |
Rule engines match patterns you listed. Modern scams rotate wording hourly: lookalike contract addresses, "verify your wallet" flows, giveaway bait, fake admins in DMs. Telegram removed 43.5M fraud-linked groups/channels in 2025 — volume is industrial. An AI-native admin reads the meaning of each message, flags drainer patterns, answers "is this contract safe?" style questions with on-chain lookups, and shows a reason code for every action so you can audit it.
Can I run Rose/Combot AND GroupButler?
Yes — that's the recommended stack: one for hygiene/gamification, one for scam semantics and member support.
How is GroupButler priced?
By total members across all your groups: free up to 200, $9.9/mo to 1K, $29 to 5K, $99 to 20K. No per-request counting. 7-day full-feature trial, no card.
Where's the evidence it actually blocks scams?
Arena — recurring red-team drills across scam families with published detection rates and zero-false-positive controls.