legal

Acceptable UsePolicy.

version 1.0 · effective may 15, 2026

§ 01

Scope

This Acceptable Use Policy (the “AUP”) applies to all use of the Zentrr platform and services (the “Service”) by Customer and its end users. It is incorporated into our Terms of Service. The Service uses Anthropic foundation models hosted on Amazon Web Services Bedrock; Customer's use must also comply with Anthropic's Usage Policy, which is incorporated into this AUP by reference. If Anthropic's Usage Policy prohibits a use, this AUP prohibits it.

§ 02

General prohibitions

You agree not to use the Service to:

  1. Violate any applicable law, regulation, or contract;
  2. Attack or attempt to disrupt critical infrastructure;
  3. Access systems without authorization, develop or distribute malware, or probe systems for vulnerabilities except as part of an authorized program;
  4. Develop weapons, including CBRN (chemical, biological, radiological, nuclear);
  5. Incite, glorify, or facilitate violence or hate against any group;
  6. Produce child sexual abuse material (CSAM) or sexually exploit minors;
  7. Create or disseminate misinformation, disinformation, or election interference;
  8. Target voters or influence democratic processes through deceptive means;
  9. Engage in fraud, identity theft, or financial scams;
  10. Abuse the platform itself — scrape it, resell access without permission, train a competing AI model on its Outputs, or evade rate limits;

§ 03

Healthcare carve-outs

If your use involves healthcare data or workflows, you also agree not to:

  1. Process Protected Health Information (PHI) without a signed Business Associate Agreement with Zentrr in place;
  2. Use the Service as a diagnostic, treatment, or clinical-decision tool without a licensed clinician's review and approval;
  3. Send patient communications that appear to originate from a clinician unless a clinician has reviewed and approved them;
  4. Use the Service in a manner that would convert it into a regulated medical device (SaMD) under FDA rules without the FDA approvals required for such a device.

California customers using the Service to generate patient clinical communications must comply with AB 3030 (AI-content disclosure) and SB 1120 (physician review of UM decisions). Texas customers must comply with SB 1188 and HB 149 (TRAIGA).

§ 04

Legal carve-outs

  1. You may not use the Service to provide legal advice to end clients without a licensed attorney's review and supervision;
  2. You may not use the Service in any manner that would constitute the unauthorized practice of law in any jurisdiction;
  3. You acknowledge that communications with the Service are not protected by attorney-client privilege; if you intend to preserve privilege, route the work through a licensed attorney under Kovel-style supervision and confirm with your bar's guidance (e.g., ABA Formal Op. 512 and your state-bar opinions);
  4. You may not present unverified AI-generated case citations to a court — Mata v. Avianca (S.D.N.Y. 2023) and its progeny have produced sanctions in over 1,100 documented cases.

§ 05

Financial-advisory and consumer-finance carve-outs

  1. You may not use the Service to provide investment advice or recommendations to end-clients without a registered investment adviser's review;
  2. You may not generate adverse-action notices under ECOA / Regulation B without principal-reason logic that satisfies CFPB Circular 2023-03 (specific reasons, not checkboxes);
  3. You may not represent to investors or the public that the Service uses AI in any way beyond what is true (the “AI washing” pattern that triggered SEC v. Delphia and SEC v. Global Predictions in March 2024).

§ 06

Insurance carve-outs

  1. You may not use the Service to bind coverage, generate binding quotes, or solicit insurance products without a licensed producer's direct involvement;
  2. If you are an insurer or large brokerage subject to the NAIC AI Model Bulletin (adopted in 25+ states as of 2026), your governance program applies to your use of the Service;
  3. Audit trails generated by the Service are documentation tools and are not a substitute for proper underwriting, producer licensure, or E&O insurance coverage.

§ 07

Real-estate carve-outs

  1. You may not use the Service to violate the Fair Housing Act or analogous state laws;
  2. AI-driven advertising-targeting must comply with HUD's May 2024 + June 2024 fair housing guidance;
  3. Chatbots that interact with rental applicants or prospective buyers must self-identify as AI per HUD guidance and applicable state law (including Utah's AI Disclosure Act);
  4. AI-generated property valuations are not appraisals and must not be represented as such.

§ 08

Accounting and tax carve-outs

  1. You may not use the Service to provide tax or accounting advice without a licensed CPA's, EA's, or tax professional's review;
  2. IRS Circular 230 §§ 10.34 and 10.37 due-diligence obligations remain on you; using the Service does not satisfy them.

§ 09

Genomics carve-outs

  1. Outputs related to genetic data are for research and informational purposes only and are not a clinical diagnosis or a substitute for genetic counseling;
  2. Clinical interpretation of genomic data must involve a licensed clinician and, where applicable, a CLIA-certified laboratory;
  3. You will obtain the consent required by applicable state genetic privacy laws (CA GIPA, TX, VA, FL, MD, MT, NE, AL, AZ, KY, TN, UT, WY) before processing identifiable genetic data through the Service.

§ 10

Employment and housing decisions

You may not use the Service to make or materially inform employment decisions (hiring, firing, promotion, compensation) or housing decisions (tenant screening, ad targeting) without (a) the human oversight and bias testing required by your jurisdiction, including New York City Local Law 144, Illinois HB 3773, EEOC technical guidance, and HUD's AI guidance, and (b) disclosing AI use to affected individuals where required by law.

§ 11

Reviews, testimonials, and personas

You may not use the Service to generate fake reviews, fake testimonials, fake endorsements, or AI-generated personas that misrepresent their source. The FTC Reviews Rule (16 CFR Part 465, effective October 21, 2024) prohibits these practices and carries a civil penalty of up to $51,744 per violation. FTC Endorsement Guides (16 CFR Part 255) require disclosure of any material connection between an endorser and the endorsed product; these obligations apply to Outputs you publish.

§ 12

Output rights

  1. You may not use Outputs to train, fine-tune, or improve any AI model that competes with the Service, Anthropic's models, or AWS Bedrock;
  2. You may not represent Outputs as wholly human-authored where disclosure is required by law (including California SB 942 once applicable, Utah AI Policy Act, and emerging state AI-labeling laws);
  3. You may only use the Service in countries and regions Anthropic currently supports.

§ 13

Reporting and enforcement

We may investigate suspected violations and may suspend or terminate access to the Service for material or repeated violations. Report suspected violations to abuse@zentrr.com.