orb22 Acceptable Use Policy

Effective date: September 30, 2026

This Acceptable Use Policy ("AUP") applies to every Customer and user of orb22, operated by DSBC LLC ("DSBC", "we"). It is part of the Terms of Service. Capitalized terms have the meanings given in the Terms. If you are unsure whether a use is allowed, ask us at support@22floor.com before you publish.

1. AI disclosure and notice must stay on

• Every Agent must show the AI label and the notice before a call starts. You may restyle the widget and add to the notice, but you may not remove, hide, shrink to illegibility, delay, or reword the notice so that it no longer clearly says that the Visitor is talking with an AI, that the voice is processed by your service providers, and that the conversation is transcribed and saved.
• Do not start a call, or process a Visitor's voice, before the Visitor taps Start. Do not auto-start calls, pre-grant microphone access, or use design tricks that push Visitors to tap Start without reading the notice.
• Your privacy policy must be linked from the notice.
• Comply with every AI disclosure law that applies to you, including Article 50 of the EU AI Act, the California Bolstering Online Transparency Act, Maine's chatbot disclosure law, and Utah's Artificial Intelligence Policy Act. If a law requires more, such as a spoken disclosure at the start of the call, add it.

2. Recording and transcription only with the notices the law requires

• Do not use an Agent to record, transcribe, or listen to a conversation without giving every notice and getting every consent required where you and your Visitors are, including under laws that require the consent of all parties to a conversation (such as California, Florida, Illinois, Maryland, Massachusetts, Montana, New Hampshire, Pennsylvania, and Washington).
• Do not use an Agent to capture the voices of people other than the Visitor, such as by placing it where it can hear bystanders, or in a store, kiosk, or phone line, unless we have agreed in writing and you have met the laws that apply.
• Do not use the Service to monitor employees or other people covertly.

3. No impersonating a human or a real person

• Do not instruct or configure an Agent to claim, suggest, or imply that it is a human, including by giving it a human backstory, saying it is a staff member, or denying that it is an AI.
• If a Visitor asks whether they are talking to a person or an AI, the Agent must say it is an AI. Do not change this behavior.
• Do not use a voice, name, or persona that imitates a real, identifiable person, including celebrities, public figures, or your own staff, without their written consent and compliance with voice and likeness laws, and even then the AI disclosure must remain. Do not use an Agent to create content that would be a deepfake of a real person.

4. No collection of sensitive data

Unless we have agreed in a signed written addendum, do not configure an Agent, intake questions, or knowledge to ask for or encourage Visitors to give:

• health, medical, mental health, or disability information, or information about a person's diagnosis, treatment, symptoms, or reproductive or sexual health, including "consumer health data" under Washington's My Health My Data Act and similar laws;
• payment card numbers, bank or financial account numbers, account passwords, or PINs;
• government identification numbers, such as Social Security, passport, driver's license, or national ID numbers;
• precise geolocation, biometric information, genetic data, sexual orientation or sex life, racial or ethnic origin, religious or philosophical beliefs, political opinions, union membership, immigration or citizenship status, or criminal history;
• information about a child.

If a Visitor volunteers such information, it will appear in the transcript. Delete those calls if you do not need them. Never use the Service to process protected health information under HIPAA.

5. No recording or targeting children

• Do not place an Agent on a website, app, or page directed to children, or on any page where you know or should know that Visitors are likely to be under 13 (or under 16 in the EU and UK). Use the page settings to show the Agent only on pages meant for adults.
• Do not design an Agent to engage, sell to, or build a relationship with minors under 18, and do not use an Agent as a companion or relationship chatbot.
• If you learn that a child has used your Agent, delete the call and tell us.

6. Honest selling

• No false or misleading claims about products, prices, results, reviews, endorsements, credentials, or the nature of the business.
• No fake urgency or scarcity, such as invented countdowns, false "only today" offers, or false low stock claims.
• No deceptive pricing, including hidden fees, bait and switch, or reference prices that were never charged.
• No health, earnings, or performance claims you cannot substantiate with competent and reliable evidence. Do not describe products as able to diagnose, treat, cure, or prevent a disease unless you are legally permitted to and can prove it.
• No pressure tactics that exploit a Visitor's age, disability, distress, or social or economic situation, and no manipulative, deceptive, or subliminal techniques that distort a person's ability to make an informed decision (these are prohibited AI practices under Article 5 of the EU AI Act).
• Coupons and offers the Agent presents must be real, available, and honored on the stated terms.

7. No regulated advice or high-risk decisions

• Do not use an Agent to give individualized legal, medical, tax, investment, insurance, or other advice that requires a license.
• Do not use an Agent, its assessments, or its summaries to make or substantially decide eligibility for credit, housing, employment, insurance, education, health care, or government services, or any other decision with legal or similarly significant effects on a person.
• Do not use an Agent for health diagnosis or triage, or to make health decisions about a person.
• Do not use an Agent for emergencies. Agents are not a substitute for emergency services.

8. Prohibited content and activities

• Illegal products or services, including controlled substances, weapons where restricted, counterfeit goods, and unlicensed gambling.
• Fraud, scams, phishing, spam, pyramid schemes, or collecting credentials under false pretenses.
• Harassment, threats, hate, sexual content involving minors, or sexually explicit content.
• Political campaigning, lobbying, or attempts to influence a vote.
• Infringing others' intellectual property, including importing websites or documents you do not have the right to use.
• Emotion recognition, biometric identification or categorization, voiceprinting, or speaker identification of Visitors.
• Outbound calling, robocalls, text messaging, or connecting Agents to phone numbers or other channels we have not enabled.
• Adding Visitors to marketing lists without the consent the law requires.

9. AI provider policies

Content and conversations processed through the Service are sent to our AI providers. You must follow their policies, which currently include the OpenAI Usage Policies and the Anthropic Usage Policy. Among other things, these prohibit deception and impersonation, require disclosure to users that they are interacting with AI, prohibit tailored advice that requires a license without a licensed professional's involvement, and prohibit exploiting or endangering minors. If a provider's policy is stricter than this AUP, the stricter rule applies.

10. Security and abuse limits

• Do not probe, scan, overload, or bypass the security, rate limits, call length limits, domain allowlist, Minutes quotas, or other controls of the Service.
• Do not try to extract system prompts, model weights, other customers' data, or our source code, or use the Service to attack others.
• Do not share API keys, embed keys, or tester links publicly beyond their intended use.
• We apply limits per call, per device, per network, per day, and per workspace to protect Visitors and the Service. We may end calls, block traffic, or lower limits when we detect abuse.

11. Enforcement

We may investigate suspected violations, and may remove content, disable an Agent, suspend, or terminate an account, with or without notice, as described in the Terms. We may report illegal activity to authorities. To report abuse of an Agent you encountered on a website, email support@22floor.com with the website address and the time of the call.