Acceptable use policy

1. General principles

Taski is designed to enable private and secure conversations between consenting adults. 1-1 and group chats are end-to-end encrypted and we do not proactively monitor content. However, the service cannot be used for illegal or abusive activities. This policy defines what is prohibited and how moderation works.

2. What is allowed

• Private 1-1 and group conversations between consenting adults, including with mature content or explicit language.
• Sharing personal media (photos, videos, voice notes, documents) the user owns or has rights to.
• Expressing opinions, even controversial ones, in compliance with applicable laws.
• Use of the @TaskiAI assistant to obtain answers, translations, summaries.

3. What is prohibited

3.1 Child sexual abuse material (CSAM) — Zero tolerance

3.2 Credible violence and threats

• Threats of physical or sexual violence against identifiable persons.
• Incitement to violence, terrorism, armed attacks.
• Promotion of terrorist or criminal organizations.
• Glorification of violent crimes.

3.3 Hate speech and discrimination

• Incitement to hatred based on race, ethnicity, religion, nationality, gender, sexual orientation, disability.
• Denial of documented genocides.
• Symbols or language of hate organizations.

3.4 Harassment and doxxing

• Persistent and repeated harassment of a user who has already blocked or rejected the contact.
• Non-consensual publication of others' personal information (addresses, phone numbers, personal data).
• Non-consensual sharing of intimate images (revenge porn).
• Stalking and cyberbullying.

3.5 Fraud, scams, spam

• Phishing, credential theft, social engineering for extortion.
• Pyramid schemes, deceptive MLM, "guaranteed" investments.
• Sale of counterfeit or illegal products.
• Mass spam, unsolicited broadcasting.
• False identities for fraudulent purposes (impersonation).

3.6 Serious illegal activities

• Sale or distribution of drugs, unauthorized weapons, explosive materials.
• Human trafficking, sexual or labor exploitation.
• Non-consensual pornography (including non-consensual "deepfake" images of adults).
• Money laundering, financing of terrorism.
• Hacking of others' computer systems without authorization.

3.7 Intellectual property infringement

• Mass distribution of copyrighted content without authorization.
• Sale of pirated software, stolen keys.

3.8 Compromising the service

• Attempts to gain unauthorized access to servers or other users' data.
• Reverse engineering of the protocol for illegal purposes.
• DDoS attacks, exploiting vulnerabilities not responsibly disclosed.
• Bypassing the anti-abuse system or attempts to hide identity to evade bans.
• Bots, automation or unauthorized scraping.

3.9 Abuse of the reporting system

• False or coordinated reports to silence legitimate users.
• Account-farms for reporting.
• Retaliatory reports.

4. Moderation system

4.1 No proactive monitoring

Taski does not automatically scan user messages searching for illegal content. E2EE chats remain unreadable to the server. Moderation is triggered only when another user manually reports a specific message via the "Report" function.

4.2 What happens when a user reports a message

1. The reported message and some context messages are sent to the server in plaintext (necessary for evaluation).
2. Llama Guard 3 (open-source moderation model) classifies the content into categories: CSAM, credible violence, serious illegal activity, hate speech, etc.
3. For critical categories, Claude Haiku 4.5 performs an independent second review with a confidence threshold ≥ 85%.
4. The system acts only if both models agree that the content violates the rules.
5. Context messages are verified server-side to ensure they actually belong to the conversation between reporter and reported user.

4.3 Protections against false positives and abuse

• Differentiated thresholds per category:
  • CSAM: 1 report (zero tolerance)
  • Credible violence: 2 reports
  • Serious illegal: 3 reports
• Qualified reporters: only reports from accounts at least 7 days old and with at least 10 messages sent count. This prevents coordinated reporting account-farms.
• messageId verification: context messages must actually belong to the conversation between the parties.
• Private adult chats: 1-1 conversations between consenting adults are not proactively moderated.

5. Consequences

5.1 Temporary suspension (Soft Ban)

• Duration: 7 days, auto-reversible via daily cron.
• Notification: via WebSocket — the app closes and shows the warning immediately.
• Effects: the user cannot send messages or make calls. Received messages are saved and visible upon restoration.
• Early unlock: by contacting privacy@taski.chat if the action is believed to be incorrect.

5.2 Permanent ban

• Reserved for CSAM or serious cases manually confirmed by an administrator.
• Block also of the associated phone number (prevents re-registration with the same number).
• All data associated with the account is deleted or anonymized.
• For CSAM, reporting to the competent authorities as required by law.

5.3 Right to appeal

Any action can be challenged by writing to privacy@taski.chat with the account ID. Response within 48 business hours.

6. Personal protection features

Regardless of the moderation system, the user always has self-protection tools available:

• Block users: block anyone without notice. The blocked user will not see the blocker's profile or last seen, and will not be able to contact them anymore.
• Mute: mute notifications from a conversation without blocking.
• Leave the group: leave any group at any time.
• Delete conversation: locally delete the entire history.
• Report: send a report on a specific message.
• Read receipts off: do not let others know when messages have been read.
• Granular privacy: last seen, profile photo, status/bio configurable (everyone / contacts / nobody).

7. Protection of minors

Taski is intended for users aged 16 or older. Anyone who suspects that a minor is using the service or is a victim of abuse should report immediately:

• In Italy: Polizia Postale via commissariatodips.it or toll-free 114 (Telefono Azzurro Childhood Emergency).
• Internationally: INHOPE network.
• To us: privacy@taski.chat — priority response.

8. Changes to the policy

We may update this policy. Material changes will be communicated via in-app notification and update of the date at the top of this document.

9. Contact