FAQ - eGuardian

Which AI features are available to use?

eGuardian offers natural language conversations, AI image generation, context-aware responses, and advanced spam/content detection powered by Gemini AI. Members can chat naturally with the bot, request image generation, and get intelligent responses 24/7.

What exactly are Defenders?

Defenders are regular members with one special right, to vote against SPAM or inappropriate content. Instead of solely relying on admins, trusted members can participate in democratic moderation by voting on whether to remove messages or take action against offenders. All votes and actions are transparently logged for accountability. Admin can assign Defenders status to any trusted member. Defenders are not admins and have no extra privileges beyond voting. This doesn't affect normal telegram voting mechanisms.

Will the bot delete legitimate messages?

Advanced AI analysis + fully configurable detection = minimal false positives. You control exactly what gets detected and can whitelist specific users without giving them admin rights.

What data do you collect from my group?

We only collect data necessary for moderation: violations, removed messages, and action logs. We never store private conversations or user personal information beyond what's needed for bot functions.

Can I disable specific detection types?

Yes! Complete control per group. Enable or Disable detection types or AI detections as needed. You're in charge. Fully customizable to fit your group's needs.

Are there manual moderation commands?

Yes! Admins can use commands to manually remove users, mute members, ban users, or trigger custom actions. All manual actions are logged transparently and appear in the action history for accountability and visibility.

Is there a cost or free trial?

eGuardian is completely free! All features available for all groups. No hidden costs, no premium tier. We believe group protection should be accessible to everyone.

What happens if the bot goes offline?

The bot runs on enterprise-grade servers with 99.9% uptime guarantee. In the unlikely event of downtime, we notify admins immediately. Your group remains safe and all settings are preserved.

How do I customize the spam detection prompt?

In the Group Management page, you can customize the spam detection context. The AI will use your custom prompt to identify spam. Here's an example of a detailed prompt for a cryptocurrency project group:

Example Prompt:

You are monitoring a Telegram group dedicated to the DIGICOIN (DIGI) cryptocurrency project.

Project Information:
- Legitimate project: DIGICOIN or DIGI (Solana-based token)
- Official contract address: AUdUEc98MGfEHJiJfCgMaW8gKdcfNDio8BFzGKBwjztC
- Blockchain: Built on Solana network

SPAM/SCAM Detection Rules:
• Flag as SPAM if the message contains:
  - Advertisements for unknown/suspicious crypto projects or memecoins
  - Contract addresses other than the official DIGICOIN address
  - Pump and dump schemes or "get rich quick" promotions for unknown tokens
  - Phishing links or suspicious URLs requesting wallet connections
  - Fake giveaways, airdrops, or "double your crypto" offers
  - Impersonation of team members or official announcements
  - Promotional content for unverified/scam tokens and projects
  - Nudity, sexual content, or inappropriate imagery

ALLOWED Content:
• Everything else is allowed, including:
  - All language types (profanity, criticism, frustration, strong opinions)
  - DIGICOIN discussions, price talk, and technical questions
  - Any cryptocurrency discussions (Bitcoin, Ethereum, Solana, all major coins)
  - General conversations, memes, jokes, and community interactions
  - Political discussions, personal opinions, and debates
  - Market analysis, trading discussions, and investment talk
  - Off-topic conversations and casual chat

Decision Guidelines:
- Only block content that is clearly SCAM or NUDITY
- Everything else = CLEAN (be very permissive)
- When uncertain between SPAM/CLEAN → always choose CLEAN
- Focus ONLY on financial scams and inappropriate imagery
- Allow all other content regardless of topic, tone, or language
- Prioritize catching contract address scams for unknown projects and nudity

Frequently Asked Questions