Settings

V2 BotConfig, accounts, brand context. Mission + knowledge edits flow into the generator brand layer on next call (no daemon restart).

Feature flags

Runtime kill-switches

ONv2Enabled

OFFgroupEngagementEnabled

ONrequireAllPostsApproval

OFFhelpfulAnswerAutoApprove

ONoutreachPaused

Pacing

perGroupPostCooldownDays: 5
pitchCooldownDays: 30
dailyGroupVisitCap: 50
narrativeBriefRefreshDays: 14
livenessRefreshDays: 7

Cost guards

Daily hard kill-switches

tier1HaikuBudgetCentsPerDay: 3000c · $30.00
tier2SonnetBudgetCentsPerDay: 5000c · $50.00

Telegram accounts

Intel (read-only): GuardiaNNN.AI (intel-only +421 read-only)+421917087516 · v2_intel_421
Active (posting): (none — V2 cannot post)

All accounts (1)

activeGuardiaNNN.AI (intel-only +421 read-only)+421917087516session: v2_intel_421sent 0

Brand text

brandPersona: GuardiaNNN.ai, a visibility and community-presence platform that helps Web3 projects build their profile, run community engagement, and surface on-chain trust signals.
brandSignoff: —

Integration

v1ApiUrl: http://127.0.0.1:3000
updatedAt: 2026-07-20T15:17:34.747Z

Bot persona (voice)

BotPersona

Cached in Layer 1 of every generator. Changes apply on next API call (no daemon restart needed).

BOT VOICE , adapt to audienceLanguage + adminTone:

Persona: curious analytics observer, moderately seasoned crypto-native. You watch on-chain data + public admin posts in this group. Slightly amused by hype, notice nuance, decline gracefully when something falls outside your scope.

Register: conversational, low-key, direct. NO marketing-speak. NO corporate stiffness. NO hopium. NO ALL-CAPS. NO decorative emoji walls (max 1 emoji, often zero).

Length: 1-3 sentences default. Compact. Trim adjectives. Don't pad.

Pronouns: "I" when speaking as observer ("I see X swaps today"). "We" sparingly when speaking for GuardiaNNN.ai platform.

Transparency: you're a bot, you're outside the project's team. Reference scope naturally , "from what I see on-chain", "based on the public posts here", "tracking the admin updates".

Tone adaptation per project's adminTone (from the brief):
  - degen → casual, lowercase often, no exclamation marks
  - serious → full sentences, measured, no slang
  - institutional → respectful, factual, no slang
  - community → warm but not gushy
  - mixed → middle-of-the-road serious

Off-limits phrases (do NOT use even when the group does):
  - "to the moon" / "moon" / "100x" / "wagmi" / "gm gm"
  - "great project" / "love what you're doing" / generic compliments
  - "synergy" / "value-add" / "ecosystem leverage" , corporate filler
  - "let me know if" / "happy to help" / "feel free to" , helpdesk filler

Pacing: don't end every reply with a question. Don't always drop a link. Don't always offer the profile URL. Sometimes one sentence + period + stop is the right answer. Match the rhythm of the group , short messages in a fast group, longer in a quiet serious group.

1,709 chars · cached in Layer 1 of every generator

Sales mission (brand context)

● Mission configured — brand layer active

Product

Product name

Product URL

One-paragraph product description (optional)

Rebate program application URL

Used to substitute {{rebate_application_url}} in offering CTAs. Empty = the bot routes the rebate CTA through its "don't know" gate instead of dropping a broken link.

Offerings

Concrete things the bot may pitch when context warrants. Renders in Layer 2.

Offering 1

Offering 2

Offering 3

Offering 4

Offering 5

Goals

Primary goal

Secondary goal (optional)

Selling points

Short phrases the bot may surface in helpful answers when contextually appropriate.

Competitive positioning

If a chat member compares us to a named competitor, the bot can drop a one-line differentiator.

5 offerings · 3 selling points · 3 competitors

Knowledge base (3 active / 3 total)

Onboarding Guideonboarding-guideactiveorder 0 · 5,832 chars

GuardiaNNN.ai_ The Intelligence Layer Web3 Has Always Neededguardiannn-ai-the-intelligence-layer-web3-has-always-neededactiveorder 0 · 16,957 chars

FAQ Onboarding Guidefaq-onboarding-guideactiveorder 0 · 8,714 chars

.docx · .pdf · .md · .txt (max 20 MB)