Using /ask and /trust

/trust

Shows the public trust summary for the current project's community: confidence-gated scores, official links, and an AI-generated public-safe summary.

• Works in verified project groups.
• Output is identical to the project's public Trust Page.
• "Not enough data yet" means confidence is too low to publish a score — see Confidence levels.

/ask

Requires Starter+ full Telegram access. /ask is not available on Free Lite. On paid plans, /ask queries an AI assistant grounded only in the owner-approved Knowledge Base — facts, FAQs, and chunks extracted from uploaded PDFs and URLs that an owner or manager has approved.

• Answers reflect approved facts only. If the answer isn't there, the bot says so.
• The assistant prepends @username when replying in groups.
• Stale or superseded facts are filtered out automatically (Knowledge freshness).
• Rate limits apply per group to keep things fair.
• It is not financial advice, ever.

Reply modes (Off / Drafts / Ambient)

Owners pick how the AI behaves from the Owner Control Center:

• Off — bot never replies with AI content.
• Drafts — AI generates a draft for every triggered reply; nothing posts publicly until you approve it in the Owner Review Queue.
• Ambient — bot quietly listens and only auto-replies to high-confidence questions backed by approved facts. Risky or low-confidence questions still route to drafts.

Why we are conservative with AI

Crypto is high-stakes. We would rather say "I don't know" than fabricate. If the assistant refuses or hedges, that is a feature, not a bug.

See also

• All Telegram commands
• Reporting suspicious activity
• Terms of Service