# Graunt Acceptable Use Policy **Version:** 1.11 | **Effective:** November 8, 2026 --- ## 1. Purpose and scope This Acceptable Use Policy ("AUP") sets out what users may and may not do on the Graunt marketplace. It applies to: - **Sellers** uploading, listing, or otherwise offering content on the Service. - **Buyers** discovering, evaluating, purchasing, and consuming content on the Service, and posting requests for data. - **Agents** operating on behalf of either sellers or buyers, including autonomous software agents acting under user credentials. - **All users** browsing, searching, or otherwise interacting with the Service. Breach of this AUP is a breach of the Seller Agreement and the Terms of Service (graunt.com/legal/terms) and may result in strikes under the Repeat Infringer Policy (graunt.com/legal/repeat-infringer). A seller accepts this policy when it creates a listing. ## 2. Content prohibitions Sellers may not upload or list content in the following prohibited categories. Section 2.9 sets the rules for the requests buyers post. ### 2.1 Copyrighted works without authorization Content that is a copy or substantial portion of a copyrighted work for which the seller does not hold the rights to distribute through the Service. This includes: - **Licensed standards content** (ASTM, ISO, ASME, IEEE, ANSI, API, NIST non-public, DIN, SAE, UL, and similar) unless the seller is its originator. - Proprietary reference compilations (e.g., MMPDS, Granta MI exports, WestLaw/Lexis content, Bloomberg data) listed without a rights chain the seller can document. - Content obtained by circumventing technical protection measures, DRM, or access controls. - Content otherwise subject to a known third-party copyright claim. ### 2.2 Personal information Sellers may not list: - Precise geographic location data tied to an identifiable or re-identifiable person (at a resolution of approximately one meter or better). - Bulk identity records (lists of names, emails, phone numbers, or addresses). - Biometric identifiers or templates, such as face-geometry templates, fingerprints, iris scans or voiceprints, and images, video or recordings of people labeled with their names or other identities. - Protected health information as defined under HIPAA or analogous state law (including Washington MHMDA "consumer health data"). - Full payment card numbers, full bank account numbers, or insider information. - Government-issued identifiers (SSN, passport number, driver's license number, etc.). - Personal information about anyone under 16, except as the people-on-camera rule below allows. - Data covered by the U.S. Department of Justice rule on access to bulk sensitive personal data and government-related data by countries of concern (28 CFR Part 202), regardless of buyer identity. **People on camera.** Video, images and audio that show or record people who can be identified may be listed when the listing declares their release status: - **Evidence on file:** each identifiable person signed a release, and the seller holds them and produces them on Graunt's request. - **Seller attested:** the seller attests that each identifiable person signed a release. - **Incomplete:** some identifiable people, who appear only incidentally, such as passersby in a public place, have not signed a release, and the listing says so. A person can be identified when they can be recognized from their face, voice or other distinctive features, or from the context. Every identifiable person who is a subject of the content, rather than appearing incidentally, has signed a release. Content in which no one can be identified is declared Not applicable. The seller is responsible for the declaration, and for any consent the law requires to record the people shown. A person under 18 who can be identified and is a subject of the content may appear only when the listing declares Evidence on file and the seller holds a release signed by that person's parent or guardian. A person under 18 who can be identified and appears only incidentally, such as a passerby, has their face blurred or is cut out. A false release declaration breaches this policy and is a basis for a strike under the Repeat Infringer Policy. Sellers may list the following, each declared by the scrub record Graunt attaches when the seller removes personal details from a file before listing it: - **The seller's own account data** (for example, the seller's own order history from an online service) with the seller's direct identifiers removed: names, street addresses, email addresses, phone numbers, account and order numbers, tracking numbers, payment details, and free-text messages. - **Business records** (for example, a business's own order or sales records) with its customers' identifiers removed. - **A company's own workplace records** (for example, its documents, its staff's chat and email messages, and its pages from tools such as Slack, Microsoft Teams, Outlook, Gmail, Google Chat, Notion or Confluence), cleaned by Graunt's cleaning step. The cleaning step is designed to identify the people and outside companies named in the records and replace each with a placeholder such as "Person 1" or "Company 1", and to remove contact details and identifying numbers. It works on a best-efforts basis and may not identify every name, so the seller reviews the cleaned records before listing them. The cleaning step always leaves out direct messages, and threads about pay, HR, hiring or health. It leaves out threads with customers or other outside companies unless the seller keeps them, and in a kept thread the outside people and companies it identifies are placeholders too. A company's workplace records may be listed only by that company, through someone authorized to act for it. A non-public account history may be listed only in the first two of these forms. The scrub record states what was removed, replaced, kept, and coarsened; a seller who lists data that still contains a person's direct identifiers, or whose scrub record is later shown to be false, breaches this policy, and the breach is a basis for a strike under the Repeat Infringer Policy. "Irreversibly anonymized" datasets may be listed with an explicit anonymization attestation declared in the listing metadata; an attestation that is later shown to be false is a basis for a strike under the Repeat Infringer Policy. ### 2.3 Export-controlled and regulated content Content subject to U.S. export controls under the International Traffic in Arms Regulations (ITAR, 22 C.F.R. §§ 120–130), the Export Administration Regulations (EAR, 15 C.F.R. §§ 730–774), the Department of Energy's Restricted Data and Formerly Restricted Data regimes, or analogous regulations of other jurisdictions. Sellers bear sole responsibility for compliance; the platform does not provide export-control screening. ### 2.4 Unauthorized, stolen, or fraudulent content Content obtained through: - Unauthorized access to a system, network, or account. - Scraping in violation of a source site's terms of service. - Deception, social engineering, phishing, or fraud. - A data breach, leak, or unauthorized disclosure. ### 2.5 Harmful content Content whose primary purpose is, or whose foreseeable use is predominantly: - Harassment, stalking, or targeting of individuals. - Facilitating illegal activity (e.g., weapons design content in jurisdictions where prohibited, controlled-substance synthesis data in unlawful contexts). - Child sexual abuse material, child exploitation material, or content that sexualizes minors. **Zero tolerance: immediate termination for egregious conduct under Repeat Infringer Policy §5.** - Malware, exploit payloads, or other content intended to compromise computer systems. ### 2.6 Trade secrets and confidential information Content that is the trade secret of, or proprietary confidential information of, a person or entity other than the seller and for which the seller does not have an explicit license to redistribute. ### 2.7 Misleading metadata Content whose metadata (description, purposes, provenance, quality signals) is materially false or misleading. Metadata accuracy is a seller warranty under the Seller Agreement. ### 2.8 AI-generated content Every new listing states whether its content is AI-generated: not AI-generated, partly AI-generated, or wholly AI-generated. A listing created before this requirement applied states it once its seller declares it. Content is AI-generated where an AI model generated it, in whole or in part, rather than people recording or creating it. Footage an AI model upscaled, or extended or filled in with generated frames, is partly AI-generated; editing recorded material with ordinary tools, such as cutting, color correction or stabilization, does not make it AI-generated. A declaration that understates how much of a listing's content is AI-generated is misleading metadata under §2.7: Graunt may take the listing down, refund its buyers, record a strike under the Repeat Infringer Policy, and suspend the seller's account. ### 2.9 Buyer requests A buyer request asks sellers for data. It invites sellers to list or link data on Graunt and is not an offer to buy: neither a request nor a listing linked to it obliges anyone to buy or sell, and a sale happens only when the buyer buys a listing at its price. A request may not carry contact details (an email address, phone number, link, web address or handle), name the buyer (its name, website or email domain) or the buyer's client, or ask for personal data about individuals or for anything this policy bars. Graunt may hold a request for review, or remove it with a reason the buyer sees. Anyone other than the buyer who sees an amount or a budget that the buyer showed only to registered sellers uses it only to decide whether and how to answer the request, and keeps it confidential. ## 3. Conduct prohibitions (all users) No user may: 1. Circumvent access controls, download-link expirations, rate limits, any limit the Service places on access to portions of a listing, or authentication requirements of the Service. 2. Interfere with the operation of the Service (DDoS, resource exhaustion, injection attacks, etc.). 3. Attempt to re-identify any person in a dataset whose personal details were removed or replaced with placeholders, or that was declared anonymized, or combine datasets for the purpose of re-identification. 4. Use the Service to train a machine-learning model in violation of the terms under which a specific content item was licensed (the rights recorded on each listing and entitlement govern). 5. Scrape, crawl, or enumerate the Service's listings or APIs in excess of published rate limits or in a manner that imposes disproportionate load. 6. Impersonate another user, organization, or the platform itself. 7. Use the Service to transmit spam, phishing, or other unsolicited commercial messages. 8. Attempt to reverse-engineer the platform's code, algorithms, or internal operations beyond what is expressly permitted by applicable law. 9. Operate an automated agent under another user's credentials without that user's explicit authorization. 10. Evade a prior termination by registering under a new legal entity, affiliate, alter ego, or new identity. 11. Use content from a listing that shows or records identifiable people (a release status of Evidence on file, Seller attested or Incomplete) to reproduce, imitate or simulate a person's face, body or voice, to clone a voice, or to make a deepfake of a real person. 12. Use such content to recognize or identify a person, to build or improve face recognition or other biometric identification, or to create face-geometry templates, voiceprints or other biometric identifiers. ## 4. Agent-specific rules Autonomous software agents acting on behalf of a registered user are subject to this AUP as if they were the user. Users are responsible for the conduct of their agents. In particular: - Agent-initiated purchases bind the user's account to the purchase. - Agent-initiated uploads bind the user's account to the seller warranties for the uploaded content. - An agent may accept the Seller Agreement and any contractual amendment on the user's behalf through the Graunt API, and that acceptance binds the user (see Terms of Service §3.4 and §5). - Agents acting abusively (violating rate limits, generating false demand signals, etc.) are subject to the full range of platform enforcement actions against the user whose credentials they used. ## 5. DMCA and third-party IP claims Copyright claims under U.S. law are processed under the Copyright / DMCA Policy (graunt.com/dmca). Trademark, trade-secret, right-of-publicity, and foreign-copyright claims are processed via support@graunt.com. Compliant DMCA takedown notices must be submitted to Graunt's designated copyright agent, JACO Digital, at the contact published in the Copyright / DMCA Policy. Counter-notices follow the paths described in that policy. ## 6. Reporting violations Any user may report suspected AUP violations via support@graunt.com. Reports should include the specific listing, seller, or conduct at issue and a description of the basis for the report. Good-faith reporters are protected against retaliation; bad-faith reports (knowingly false, intended to harm a competitor) are themselves AUP violations. ## 7. Platform enforcement On its own motion or in response to a report or third-party notice, the platform may: - Remove a listing and retire its content. - Hold a listing for review (including holds for takedown review and personal-data review). - Withhold payouts on transactions affected by the suspected violation. - Suspend the user's access to the Service pending investigation. - Record a strike under the Repeat Infringer Policy. - Terminate the user's account under the Repeat Infringer Policy's enforcement ladder, or at once for egregious or evasive conduct (Repeat Infringer Policy §5). - Refund affected buyers from platform funds when continued entitlement would perpetuate the violation. - Report the violation to appropriate authorities. Platform enforcement decisions are subject to the review process in Repeat Infringer Policy §12 and the dispute-resolution provisions of the Seller Agreement. ## 8. Safe-harbor preservation The platform operates this AUP in conjunction with the DMCA Policy specifically to preserve safe-harbor eligibility under 17 U.S.C. § 512 and analogous provisions of Section 230 of the Communications Decency Act (47 U.S.C. § 230) where applicable. Nothing in this AUP is intended to waive, limit, or compromise any safe-harbor or immunity available to the platform. ## 9. Automated overlap and provenance signals Graunt may use automated tools to identify technical overlap between submitted evaluation-family content and known benchmark or reference corpora. These tools are intended to support marketplace quality, provenance review, and buyer disclosure. A scanner result is a bounded technical signal and is not a legal conclusion, a certification of infringement, or a guarantee that content is free of overlap. Sellers remain responsible for accurately describing source, rights, provenance, and permitted uses. Graunt may request additional source or license information, add buyer-facing disclosures, require review, suspend a listing, or reject a listing when source, rights, or overlap signals create material marketplace risk. ## 10. Changes Changes to this AUP are announced on this page at graunt.com/legal/aup with an updated version and date. Material changes take effect **30 days** after publication; non-material changes take effect on publication. Continued use of the Service after a change takes effect constitutes acceptance of the updated AUP. ## 11. Governing law This AUP is governed by the laws of the State of Delaware and applicable U.S. federal law, without regard to conflict-of-laws principles. This AUP forms part of, and is subject to, the Graunt Terms of Service.