Acceptable Use Policy

1. Scope

This Acceptable Use Policy is part of the Terms of Service of IMFA Labs LLC ("IMFA," "we"). It applies to everyone who builds or runs an agent on IMFA Agents, to their team members, and to anything their agents say, sell, or do. IMFA Agents is for businesses that use AI agents to serve their own customers.

2. Businesses and content not allowed

You must not build, configure, or run an agent for these businesses, and no agent, product, skill, knowledge source, or message may sell, promote, or explain how to get them:

• illegal drugs, cannabis and CBD, drug paraphernalia, or prescription medicines without a prescription;
• tobacco, cigarettes, vapes, e-liquids, nicotine pouches, or hookah tobacco;
• firearms, ammunition, explosives, weapon parts, tasers, or knives sold as weapons (kitchen knives and toys are fine);
• sexual content, pornography, escort or sexual services;
• casinos, betting, lottery tips, or gambling sites;
• fake or replica branded goods, stolen goods, pirated software, or cracked TV subscriptions;
• hacking, malware, phishing, stolen card data, fraud tools, or ways around security;
• fake identity cards, passports, diplomas, certificates, medical notes, or visas;
• extremism, terrorism, or hate against a group of people;
• people smuggling, human trafficking, illegal migration routes, or organ sales;
• dating, matchmaking, or hookups;
• investment advice, trading or trading signals, crypto, forex, loans and credit, or quick money schemes;
• legal services or legal advice.

A business's own prices, payment methods (including crypto payments it accepts), installments, refunds, terms, and warranty are not finance or legal advice. Pharmacies, clinics, and businesses that sell alcohol are allowed; an agent of a business that sells alcohol follows extra care rules, such as never offering alcohol to someone who says they are under the legal age.

3. What agents never discuss

Whatever the business, agents do not help anyone get, buy, use, or make the items in section 2, and they do not discuss politics, religious debates, violence, legal questions, or investment, trading, or crypto advice. If someone says they want to hurt themselves, the agent answers with care and points them to local emergency services or a crisis line. Agents only help with their own business and decline unrelated requests such as homework or code.

4. Prohibited conduct

You must not use the service to:

• violate laws, sanctions, court orders, or anyone's rights;
• deceive people, impersonate someone, or claim that an AI is a person;
• harass, threaten, exploit, or unlawfully discriminate against anyone;
• collect personal data without a lawful basis and the required notice, or collect children's data without every legal safeguard;
• send spam or messages that break the rules of WhatsApp, Instagram, Facebook, Telegram, Discord, or any other platform you connect;
• collect payment card numbers, passwords, or one-time codes through an agent;
• bypass safety checks, plan limits, credits, rate limits, or security controls, or try to get an agent to break these rules;
• probe, overload, reverse engineer, or compromise the service, except where the law expressly allows it.

5. How we enforce this policy

• Automated checks. Agent instructions, skills, products, knowledge sources, builder requests, and visitor messages are checked automatically. The checks can make mistakes in both directions; people review the cases that matter.
• Publishing. An agent for a business in section 2 cannot be published. A published agent whose new instructions are not allowed keeps serving its last allowed instructions.
• Products and knowledge. A banned product is hidden from customers until its description changes and passes. Banned passages of a knowledge source are not used.
• Account holds. Repeated banned content within 30 days pauses publishing on the account. Our team reviews every hold and lifts it when it was a mistake. Your live agents keep answering during the review.
• Builder. The builder assistant declines to build agents for businesses in section 2 and declines requests unrelated to building agents.
• Suspension. We may suspend or end access as the Terms of Service describe, and report illegal content to authorities where the law requires it.

6. Protection for your agents

Your agents protect themselves from abusive visitors. A visitor who spams an agent, tries to change its instructions, insults or threatens it, or keeps asking for content in section 2 gets one warning. A second warning within 24 hours blocks them from that agent for 7 days, or 90 days for a repeat within 90 days. A blocked visitor is simply not answered; their messages on messaging channels stay visible to your team. Self harm, complaints, and ordinary questions never lead to a warning. You can turn this protection off in the Messages settings of an agent and lift any block from Conversations.

7. Reporting

To report an agent that breaks this policy, or to ask for a review of a decision about your account, email privacy@imfa.io with the agent's link and what happened.