What is NebulaClaw

Each bot gets a dedicated 24/7 Agent. Three model tiers — Economy / Balanced / Premium — with semantic understanding and context memory.

Four ready-to-use role presets — Project Community, KOL / Alpha Group, Trader Community, Meme Community — or fully customize the persona.

Upload FAQs, whitepapers, or product docs. The Agent answers via semantic retrieval — no keyword hits required.

Describe rules in natural language. The AI enforces mute / kick / ban with contextual understanding — low false positives, no word lists to maintain.

One Agent can serve multiple Telegram groups. Audit logs and analytics are tracked per group.

Message volume, active users, sentiment trends, and moderation actions in a single dashboard.

Pain: 24/7 community coverage needs three-shift community managers — multiplied by languages for Web3 / cross-border projects. Holidays and late nights are exactly when scam links thrive.

Solution: One Agent replaces the coverage layer — auto-reply, auto-moderate, never offline. Your team only handles what the Agent escalates.

Pain: A single airdrop-phishing / fake-cashback message can complete the con in under 30 seconds; human mods respond in minutes — by then the damage is done.

Solution: The Agent detects and acts in seconds, closing the window before phishing spreads.

Pain: Keyword / regex dictionaries need constant updates as bad actors rotate domains and scripts. Miss one and trust erodes; over-block and users leave.

Solution: Describe intent in natural language (e.g. "invite links to non-whitelisted domains", "fake airdrop / referral-farming messages"). The AI judges by context — no dictionaries to maintain; precision scales with clarity of description.

Pain: Web3 / cross-border communities naturally mix EN / ZH / KO / TR / ID / RU / VI. Splitting groups is costly; merging leads to "misjudged because unreadable."

Solution: Auto-detects 7 languages, replies in the user's language, enforces rules across all of them — no separate groups or localized dictionaries needed.

Pain: Human mods apply rules inconsistently; follow-up appeals or disputes lack an evidence chain. DAOs and listed companies are especially exposed.

Solution: Every action goes into the Audit Log — triggered rule, context, action type, timestamp. Uniform enforcement, fully traceable — ready for internal and compliance review.

Pain: Traditional bots take actions but leave no data. Leadership can't see community health; KOL partnerships and campaign retros run on gut feel.

Solution: Built-in analytics dashboard — message volume, active users, sentiment trends, quantified moderation actions. Turns the "vibes black box" into a measurable asset.