-
Harvey AI
· Harvey AI Terms of Service
The agreement conditions the use of Protected Health Information with the Service on execution of a Business Associate Addendum. The agreement also states that the Service is not PCI compliant, which the customer acknowledges....
Why it matters: This provision establishes a prerequisite BAA execution for any use of PHI with the Service, consistent with HIPAA requirements for covered entities and business associates. The explicit PCI non-compliance acknowledgment restricts use cases involving payment card data and places contractual notice on customers regarding that limitation....
-
Zelle
· Zelle Terms of Service
The terms state that any unsolicited content, idea, or submission sent to Zelle automatically becomes Zelle's property with no payment owed, and Zelle may use, redistribute, or publicly disclose such submissions in any form for any purpose, including in new or existing products....
Why it matters: This provision asserts an automatic IP assignment from the submitter to Zelle for any unsolicited material, with no compensation obligation and no confidentiality requirement. Under this clause, Zelle may also publicly disclose such submissions at its discretion, which creates IP and confidentiality exposure for users or partners who communicate ideas outside of a formal agreement....
-
Zelle
· Zelle Terms of Service
In jurisdictions where full exclusion of consequential or incidental damages is not permitted, the terms cap the Zelle Network's total financial liability to any user at $100.00, to the extent permitted by applicable law....
Why it matters: This provision establishes a $100.00 ceiling on the Zelle Network's financial liability in states where broader damage exclusions cannot be fully enforced, which applies to claims arising from site use, content inaccuracies, and related matters. The practical scope of this cap depends on state law and is qualified by 'to the greatest extent permitted by law.'...
-
Zelle
· Zelle Terms of Service
The terms authorize Zelle to update these Terms at any time by posting revised versions on the site, with no direct notification to users required, and state that continued site use constitutes irrevocable acceptance of any changes....
Why it matters: This provision establishes that the legal terms governing site use can change without direct user notification, placing the obligation to monitor for changes on users. The characterization of continued use as 'irrevocable' acceptance is a stronger formulation than affirmative consent and may face scrutiny under electronic contract formation standards in certain jurisdictions....
-
Zelle
· Zelle Terms of Service
The terms state that Zelle makes no security representations for third-party sites linked from Zelle, and users irrevocably waive any claims against Zelle arising from interactions with those third-party sites, including the provision of credit card and personal information....
Why it matters: This provision requires users to waive, irrevocably and in advance, all claims against Zelle arising from third-party site interactions, including those involving credit card and personal data exposure. The enforceability of pre-dispute waivers of claims related to financial and personal data harm may be constrained by applicable consumer protection and financial services law....
-
Monitoring
These provisions have changed before.
Monitor includes same-day alerts, structured change summaries, and monitoring for up to 25 platforms.
-
Zelle
· Zelle Terms of Service
The terms state that Zelle will cooperate with law enforcement or court orders to disclose user identity information, and users waive all claims against the Zelle Network arising from Zelle's investigation actions or from actions taken by law enforcement as a result of those investigations....
Why it matters: This provision establishes that Zelle may disclose user identity to law enforcement based on suspected violations of these Terms, and that users waive claims arising from those disclosures or any resulting law enforcement actions. The scope of the hold harmless for law enforcement-consequent actions is broad and its enforceability may depend on applicable privacy and due process standards....
-
Zelle
· Zelle Terms of Service
The terms disclaim all express and implied warranties for the Zelle site and its content, including warranties of merchantability, fitness for purpose, and non-infringement, stating that all site use is at the user's own risk....
Why it matters: This provision establishes a comprehensive disclaimer of all warranties for site functionality, content accuracy, and security, to the fullest extent permitted by applicable law. As a site-level disclaimer, its scope is qualified by the User Service Agreement, which governs Zelle payment product terms and controls in the event of conflict....
-
Zelle
· Zelle Terms of Service
The terms state that Delaware law governs these Terms, with the exception of Section 10, which is governed by the Federal Arbitration Act, indicating the presence of an arbitration provision....
Why it matters: This provision establishes Delaware as the governing law for site-level terms and designates the FAA as the governing framework for Section 10, which by reference indicates an arbitration clause. The interaction between the FAA carve-out and Delaware consumer law creates a dual-framework structure for dispute resolution that may affect users' ability to rely on state-specific consumer protections....
-
Zelle
· Zelle Terms of Service
The terms authorize Zelle to suspend user accounts, remove content, and terminate accounts without prior notice based on Zelle's sole discretion determination that a terms violation has occurred....
Why it matters: This provision authorizes account suspension and termination without notice at Zelle's sole discretion, including during an ongoing investigation. The absence of a required pre-suspension notice or appeal process is an operationally significant feature of this enforcement framework....
-
Anthropic
· Anthropic Responsible Scaling Policy
The RSP establishes defined capability thresholds (ASL levels) that, when crossed by a model, require Anthropic to implement specified upgraded security and deployment safeguards and to document and publish affirmative risk mitigation cases....
Why it matters: This provision establishes the core operational trigger mechanism of the RSP: internally defined capability benchmarks that create mandatory internal obligations for Anthropic to upgrade safeguards and publish risk assessments before continuing deployment of models crossing those thresholds....
-
Anthropic
· Anthropic Responsible Scaling Policy
The LTBT holds authority to request external review of Risk Reports and to approve Anthropic's selection of external reviewers, with a requirement that all unredacted sections of a Risk Report be evaluated by at least one external reviewer....
Why it matters: This provision establishes a structural check on Anthropic's unilateral control over its own safety audit process by vesting reviewer selection approval authority in the LTBT, an entity distinct from Anthropic's operational leadership....
-
Anthropic
· Anthropic Responsible Scaling Policy
Version 3.4 revises the internal distribution requirement for unredacted Risk Reports from all regular-clearance Anthropic staff to a minimum of 200 Anthropic employees....
Why it matters: This provision changes the internal transparency posture for unredacted safety risk information, reducing the employee population with guaranteed access to unredacted Risk Reports from all regular-clearance staff to a defined minimum threshold of 200 employees....
-
Anthropic
· Anthropic Responsible Scaling Policy
Anthropic maintains a Noncompliance Reporting and Anti-Retaliation Policy that provides employees with channels to report potential RSP violations and protection against retaliation, including a pathway for informal inquiries about potential violations....
Why it matters: This provision establishes an internal whistleblower mechanism specific to RSP compliance, with anti-retaliation protections, providing a formal channel through which employees can raise concerns about Anthropic's adherence to its own stated safety policy....
-
Anthropic
· Anthropic Responsible Scaling Policy
Anthropic's ASL-3 deployment safeguards use a four-layer architecture consisting of access controls, real-time prompt and output classifiers, asynchronous monitoring classifiers, and post-hoc jailbreak detection with rapid response procedures, applied to interactions across Claude.ai and the API....
Why it matters: This provision describes the technical architecture through which Anthropic monitors and filters user interactions with its models at ASL-3 capability levels, establishing that both real-time and asynchronous analysis of user inputs and AI outputs will be conducted as a structural feature of deployment....
-
Anthropic
· Anthropic Responsible Scaling Policy
Partners seeking access to Claude with adjusted safeguards beyond the standard deployment must undergo an enhanced due diligence review assessed against two criteria: overall trustworthiness and the beneficial nature of their use case....
Why it matters: This provision establishes a gatekeeping mechanism for adjusted safeguard access, requiring partners to satisfy Anthropic's assessment of trustworthiness and use-case benefit before receiving modified deployment conditions, with Anthropic retaining discretion over both criteria....
-
Anthropic
· Anthropic Responsible Scaling Policy
The RSP establishes a capability threshold specifically addressing chemical, biological, radiological, and nuclear development capabilities, triggered when a model could substantially uplift CBRN development for moderately resourced state programs, with corresponding required safeguards....
Why it matters: This provision establishes that CBRN-related AI capabilities are subject to a defined threshold that triggers mandatory safeguard requirements, reflecting a specific operational constraint on model deployment in relation to dual-use technologies....
-
Anthropic
· Anthropic Responsible Scaling Policy
Anthropic discloses that in implementing the prior RSP, the company identified several instances of non-compliance with its own requirements, including delayed evaluation timelines, missing elicitation techniques, and evaluations not explicitly designed to meet stated scaling buffer requirements....
Why it matters: This provision discloses that Anthropic's adherence to its own stated RSP requirements has not been complete, with specific procedural gaps identified in evaluation timelines, methodology, and buffer assessment; the company characterizes these gaps as minimal risk and uses them as a basis for policy revisions....
-
Anthropic
· Anthropic Responsible Scaling Policy
ASL-3 security standards require multi-party authorization, hardware authentication, and temporary least-privilege access grants for model weights, with mandatory code review on production code affecting weights access, intended to reduce exfiltration risk....
Why it matters: This provision describes the access control architecture for model weights as a core ASL-3 security requirement, establishing procedural controls including multi-party authorization and hardware authentication tokens as prerequisites for any weights access....
-
Anthropic
· Anthropic Responsible Scaling Policy
Anthropic maintains a publicly updated Frontier Safety Roadmap that discloses specific AI safety goals, tracks their achievement, and replaces completed goals with updated objectives; the roadmap is a living document updated as safety milestones are reached or revised....
Why it matters: The Frontier Safety Roadmap provision establishes a public disclosure mechanism for Anthropic's AI safety objectives and their completion status, creating a trackable record of stated safety commitments and their evolution over time....
-
LlamaIndex
· LlamaIndex Privacy Policy
The policy discloses that LlamaIndex deploys third-party session-recording technologies that may capture mouse movements, scrolling, clicks, keystroke activity, and text entered into the platform including chat features and other communication tools....
Why it matters: This provision authorizes deployment of third-party session-recording tools that may capture text inputs in real time, including content entered into LLM-related chat or communication features, which may include sensitive, confidential, or proprietary information depending on how the platform is used....
-
LlamaIndex
· LlamaIndex Privacy Policy
The policy states that LlamaIndex may generate inferences or predictions about individual users' interests and preferences derived from collected personal data and platform interactions....
Why it matters: This provision authorizes LlamaIndex to create derived data profiles about users beyond the information directly collected, which may be used for targeted advertising, personalization, or other purposes described in the policy....
-
LlamaIndex
· LlamaIndex Privacy Policy
The policy states that where users access LlamaIndex services in an employment context, the company may disclose personal data including usage information to the user's employer....
Why it matters: This provision authorizes disclosure of individual user activity data to employers without specifying what categories of usage data may be shared or what notice, if any, is provided to the individual user prior to disclosure....
-
LlamaIndex
· LlamaIndex Privacy Policy
The policy states that it only covers LlamaIndex's data practices as a controller, and explicitly excludes from its scope any personal data LlamaIndex processes on behalf of enterprise customers acting as processors; individuals whose data is processed in that context must direct requests to the relevant customer rather than to LlamaIndex....
Why it matters: This provision establishes a structural limitation on the scope of data subject rights exercisable directly against LlamaIndex, creating a dependency on enterprise customers' own privacy notices and processes for end users of LlamaIndex-powered products....
-
LlamaIndex
· LlamaIndex Privacy Policy
The policy authorizes LlamaIndex to share hashed email addresses or user IDs with third-party advertising partners and to use cookies, pixels, mobile SDKs, and other tracking technologies to facilitate cross-site and cross-device targeted advertising based on user behavioral data....
Why it matters: This provision authorizes cross-context behavioral advertising using persistent identifiers shared with third-party ad networks, enabling user tracking across nonaffiliated websites and services beyond the LlamaIndex platform itself....
-
LlamaIndex
· LlamaIndex Privacy Policy
The policy states that personal data may be transferred to third parties as an asset in connection with mergers, acquisitions, joint ventures, financings, asset sales, or insolvency proceedings, including during the negotiation phase of such transactions....
Why it matters: This provision authorizes disclosure of personal data to prospective acquirers or creditors during deal negotiations and in insolvency proceedings, which may occur before any transaction is finalized and without individual user notification at the time of transfer....
-
LlamaIndex
· LlamaIndex Privacy Policy
The policy sets the minimum age for use of LlamaIndex services at 18, states the services are not directed at individuals under 18, and commits to promptly deleting personal data found to have been provided by individuals under that age....
Why it matters: The policy establishes an age threshold of 18, which exceeds the COPPA threshold of 13, and may interact with state-level minor protection statutes such as California's Age-Appropriate Design Code, depending on service classification and enforcement context....
-
LlamaIndex
· LlamaIndex Privacy Policy
The policy states that LlamaIndex has appointed Prighter Group as its EU privacy representative and designated it as the point of contact for data subjects in the EU to exercise rights including access and erasure requests....
Why it matters: Appointment of an EU representative under GDPR Article 27 is required for controllers established outside the EU that offer goods or services to EU data subjects or monitor their behavior. This provision establishes the operational mechanism for EU data subject rights requests....
-
LlamaIndex
· LlamaIndex Privacy Policy
The policy states that LlamaIndex collects share purchasing history, number and value of shares held, and holder of record names and addresses from shareholders and investors, using this data for investor relations management, transaction processing, tax documentation, and cap table administration....
Why it matters: The policy explicitly identifies investors and shareholders as a distinct data subject category and describes specific financial and transactional personal data categories collected from them, including data used for tax reporting and investment transaction processing....
-
LlamaIndex
· LlamaIndex Privacy Policy
The policy states that personal data is retained for no longer than reasonably necessary for its collection purpose, with discretion to retain longer for legal compliance, dispute resolution, or rights protection, and without specifying fixed retention periods for any data category....
Why it matters: The policy does not specify fixed retention periods for any personal data category, applying instead a general reasonableness standard with open-ended extensions for legal and business purposes, which may require evaluation against GDPR storage limitation requirements and applicable state law mandates....
-
LlamaIndex
· LlamaIndex Terms of Service
The agreement requires all disputes between users and LlamaIndex to be resolved through binding individual arbitration administered by JAMS, rather than through court proceedings. New users have 30 days from acceptance to opt out by emailing dispute@runllama.ai with their full legal name and a statement of intent to opt out....
Why it matters: This provision requires disputes to proceed through individual JAMS arbitration rather than federal or state court, and the agreement assigns to the arbitrator exclusive authority to resolve questions about the scope, enforceability, and formation of the arbitration clause itself. The clause applies retroactively to claims that arose before the user agreed to these Terms....
-
LlamaIndex
· LlamaIndex Terms of Service
The agreement requires users to waive participation in class actions, collective actions, private attorney general actions, and jury trials for all disputes arising from the Service. This waiver applies to both individual consumers and business entities regardless of the purpose for which the Service is used....
Why it matters: This provision requires disputes to proceed on an individual basis only, prohibiting consolidation of claims and precluding class-wide relief through arbitration or court proceedings. The agreement states that if the class action waiver is found unenforceable because it prevents a user from seeking public injunctive relief, that specific claim may be litigated in civil court while all other claims remain subject to arbitration....
-
LlamaIndex
· LlamaIndex Terms of Service
The agreement states that LlamaIndex owns all diagnostic, technical, and usage data collected from users' computers, devices, systems, and software, and that users assign any ownership rights they may have in such data to LlamaIndex. This data may be used for product improvement, research, usage monitoring, and sharing with third parties in de-identified or aggregated form....
Why it matters: This provision asserts that LlamaIndex holds exclusive ownership over Usage Data collected from user systems and software interactions, and authorizes its use for any lawful purpose including product development, analytics research, and third-party sharing in aggregated form. The assignment mechanism is drafted to cover any residual ownership interests users might otherwise retain....
-
LlamaIndex
· LlamaIndex Terms of Service
The agreement provides that any suggestions, recommendations, or feedback a user submits about the Service or LlamaIndex products is fully assigned to LlamaIndex, including all intellectual property rights, and LlamaIndex may use it for any purpose without attribution or compensation....
Why it matters: This provision operates as a present-tense IP assignment: upon submission of any feedback, all intellectual property rights in that feedback transfer to LlamaIndex. The clause covers not only direct suggestions but also ideas, know-how, concepts, and techniques contained in the feedback, which may encompass proprietary technical information submitted by organizational users....
-
LlamaIndex
· LlamaIndex Terms of Service
The agreement establishes that subscriptions are billed automatically on a weekly basis whenever a user's usage exceeds a defined threshold, and charges continue until the user cancels by contacting cancel@runllama.ai....
Why it matters: This provision establishes a usage-triggered automatic billing cycle on a weekly basis, meaning charges can be incurred without a separate affirmative enrollment action each billing period once the threshold is exceeded. Cancellation requires affirmative action by emailing cancel@runllama.ai....
-
LlamaIndex
· LlamaIndex Terms of Service
The agreement caps LlamaIndex's total liability for any claims at the greater of amounts paid by the user or $100 USD, and excludes all indirect, consequential, punitive, and exemplary damages including loss of data, profits, or goodwill....
Why it matters: This provision establishes a combined damages exclusion and aggregate liability cap that limits LlamaIndex's financial exposure to $100 or amounts paid, whichever is greater, regardless of the nature or extent of harm alleged. The exclusion of data loss damages is particularly relevant given that the Service processes user documents and data....
-
LlamaIndex
· LlamaIndex Terms of Service
The agreement requires users to defend and indemnify LlamaIndex and its affiliates, officers, and directors against all claims, damages, and legal fees arising from the user's access to the Service, violations of these Terms, User Content submitted, violations of third-party rights, and any use of the Service by third parties with the user's credentials....
Why it matters: This provision creates a broad user-side indemnification obligation covering not only intentional violations but also third-party use of the user's account credentials, which may include unauthorized access scenarios. The obligation covers direct, indirect, incidental, and consequential damages as well as legal fees....
-
LlamaIndex
· LlamaIndex Terms of Service
The agreement grants LlamaIndex a perpetual, irrevocable, worldwide, sublicensable license to use, reproduce, store, modify, distribute, and syndicate User Content as necessary to provide the Service, while explicitly stating that LlamaIndex does not train models on User Content and retains it in S3 storage for no more than 48 hours after processing....
Why it matters: The license granted is perpetual and irrevocable in its terms, but the agreement expressly limits its scope to actions reasonably necessary to provide and maintain the Service, and provides operational security disclosures including encryption in transit and at rest and a 48-hour S3 retention limit. The explicit statement that models are not trained on User Content is a material disclosure for AI platform users....
-
LlamaIndex
· LlamaIndex Terms of Service
The agreement prohibits users from using any content available on or through the Service, including metadata, for machine learning or AI training purposes, and prohibits use for technologies designed to identify natural persons....
Why it matters: This restriction in Section 2.2(c) explicitly prohibits a category of use that is particularly relevant for the developer and enterprise audience likely to use an AI document indexing platform. The prohibition covers not only direct content but also caption information, keywords, and other metadata associated with Service content....
-
LlamaIndex
· LlamaIndex Terms of Service
The agreement authorizes LlamaIndex to permanently terminate or temporarily suspend any user's account and Service access at any time, with or without notice, with or without cause, and without liability. Users remain bound by these Terms following termination....
Why it matters: This provision grants LlamaIndex broad discretion to terminate or suspend access without notice, cause, or liability, which may affect organizational users who depend on the Service for operational workflows. The no-refund policy for termination means users may lose prepaid subscription value upon termination....
-
Baseten
· Baseten Terms of Service
The agreement authorizes Baseten to compile deidentified, anonymized, and aggregated usage and performance data derived from Customer, Authorized User, and End User interactions with the platform, and asserts that Baseten owns all such Statistical Data and may use it for any lawful business purpose including product improvement and development....
Why it matters: This provision establishes that Baseten holds ownership rights over a category of data derived from Customer platform activity, which may have implications for Customer data governance policies and competitive sensitivity assessments, particularly where usage patterns or performance metrics could reveal proprietary operational information about Customer deployments....
-
Baseten
· Baseten Terms of Service
The agreement grants Baseten a perpetual, irrevocable, royalty-free, worldwide, transferable, and sublicensable license to use or incorporate into its services any feedback, suggestions, or enhancement requests submitted by Customer or its Authorized Users, without attribution....
Why it matters: This provision establishes that Customer organizations have no mechanism to withdraw or restrict Baseten's use of feedback once submitted, and that Baseten may sublicense such feedback, which may have implications for organizations whose feedback incorporates proprietary insights about model performance, architecture, or operational workflows....
-
Baseten
· Baseten Terms of Service
The agreement excludes Baseten's liability for all indirect, incidental, special, consequential, and punitive damages including lost profits, business interruption, and data loss costs, and caps Baseten's total aggregate liability for any claim at fees paid in the twelve months preceding the event giving rise to liability....
Why it matters: This provision establishes a financial ceiling on Customer's recoverable damages against Baseten that is limited to historical fees paid, and categorically excludes recovery for categories of harm that are likely to be most economically significant in a machine learning infrastructure context, such as business interruption and data loss....
-
Baseten
· Baseten Terms of Service
The DPA prohibits Customer from submitting to Baseten any personal data constituting Restricted Data, defined to include government identifiers, HIPAA-protected health information, biometric data, financial account credentials, payment card data, children's personal data under age thirteen, GDPR special category data, and criminal conviction data, unless a separate written agreement with Baseten expressly permits such submission....
Why it matters: This provision creates a contractual data intake restriction that Customer organizations must operationalize through data classification and platform intake controls, and allocates to Customer full contractual responsibility for any breach of this prohibition....
-
Baseten
· Baseten Terms of Service
The DPA requires Baseten to provide at least fifteen days advance notice of new Sub-Processor engagements via the Sub-Processor List or opt-in notification, and permits Customer to object in writing to privacy@baseten.co within five calendar days; if the objection cannot be resolved, the sole remedy available to Customer is termination of the relevant services....
Why it matters: This provision establishes the operational mechanism through which GDPR Article 28 sub-processor approval obligations are managed, and limits Customer's recourse upon unresolvable sub-processor objection to termination rather than continuation of services with a different configuration, which may be operationally significant for production ML deployments....
-
Baseten
· Baseten Terms of Service
Each Order renews automatically for successive one-year terms unless either party delivers written non-renewal notice at least sixty days before the end of the current annual term....
Why it matters: This provision establishes a sixty-day advance written notice requirement to prevent automatic renewal, which creates a contract management trigger for enterprise procurement calendars managing annual platform fee obligations....
-
Baseten
· Baseten Terms of Service
Following termination or expiration of the Agreement, Baseten provides a twenty-day window during which Customer may retrieve Customer Content through the platform; after this window, Baseten discontinues all use of and destroys all copies of Customer Content unless Customer has instructed earlier deletion....
Why it matters: This provision establishes a fixed twenty-day data retrieval window post-termination, after which Customer Content is destroyed, requiring Customer organizations to operationalize data extraction before the deadline to avoid permanent data loss....
-
Baseten
· Baseten Terms of Service
The agreement requires Customer to defend and indemnify Baseten at Customer's expense against third-party claims arising from Customer's use of Baseten products outside permitted parameters (Exclusions), Customer trademarks, or Customer's breach or alleged breach of the content and data warranties in Section 5.2....
Why it matters: This provision places defense and indemnification obligations on Customer for claims arising from alleged warranty breaches, including claims based on the allegation of a breach even where no actual breach is established, which creates exposure for Customer organizations in the event third parties challenge the legality or appropriateness of Customer Content or model deployments....
-
Baseten
· Baseten Terms of Service
The agreement authorizes Baseten to suspend all Customer and Authorized User access to the platform if any fees remain unpaid for more than thirty days, and to maintain the suspension until all overdue amounts are paid in full....
Why it matters: This provision establishes that Baseten may suspend access to production ML deployments for overdue payments, which may affect Customer's operational continuity and downstream End User services if billing issues arise....
-
Baseten
· Baseten Terms of Service
The agreement is governed by Delaware law and requires Customer to consent to personal jurisdiction and venue in New Castle County, Delaware courts for suits brought by Baseten, though no equivalent consent is required from Baseten for suits brought by Customer....
Why it matters: This provision establishes Delaware as the exclusive legal and procedural forum for claims brought by Baseten against Customer, while the venue consent is unilaterally directed at Customer, which may affect the practical accessibility of litigation for Customer organizations not based in Delaware....
-
Stash
· Stash Terms of Use (Superseded URL)
The terms authorize Stash to amend the Terms of Use at any time, with notification delivered via website posting, email, push message, text, or mail. Continued use of the platform after receiving such notification, without submitting a written objection, constitutes binding acceptance of all amended terms....
Why it matters: This provision establishes that any amendment to the Terms of Use becomes binding upon the user through continued platform use, provided notification was delivered through any of the listed channels. The mechanism does not require affirmative re-consent, placing the obligation on the user to monitor for changes and submit a written objection to avoid being bound....