Terms of Service

This document is only available in English.

These Terms of Service ("Terms") constitute a legally binding agreement between you ("you," "your," or "User") and Delph-AI ("Delph-AI," "we," "us," or "our"), governing your access to and use of the Delph-AI platform, website, and related services (collectively, the "Service").

By creating an account, accessing, or using the Service, you agree to be bound by these Terms. If you do not agree to these Terms, do not use the Service.

If you are using the Service on behalf of an organization, you represent and warrant that you have the authority to bind that organization to these Terms, and "you" refers to both you individually and that organization.

Table of Contents

1. Acceptance and Eligibility
2. Definitions
3. Description of Service
4. User Accounts
5. Subscriptions, Records, and Payments
6. Intellectual Property
7. User Content
8. Aggregated and Anonymized Data
9. AI-Specific Terms
10. Prohibited Uses
11. Disclaimer of Warranties
12. Third-Party Services and AI Providers
13. Limitation of Liability
14. Indemnification
15. Force Majeure
16. Termination and Cancellation
17. Data Retention and Deletion
18. Modifications and Notifications
19. Governing Law, Dispute Resolution, and General Provisions

1. Acceptance and Eligibility

1.1. The Service is operated by Jesús Hernández Cabañas, a sole proprietor (Régimen Simplificado de Confianza — RESICO) operating under the trade name Delph-AI, with a mailing address at Gustavo Díaz Ordaz 204, Col. Francisco Villa, Xalapa, Veracruz, C.P. 91173, México. Contact: legal@delph-ai.org.

1.2. By accessing or using the Service, you confirm that you have read, understood, and agree to be bound by these Terms and our Privacy Policy.

1.3. You must be at least 16 years of age to use the Service. If you are under 18, you represent that your parent or legal guardian has consented to your use of the Service.

1.4. The Service is designed for professional and institutional use in the context of systematic literature reviews and academic research. Individual researchers, academic institutions, pharmaceutical companies, and other organizations may use the Service in accordance with these Terms.

1.5. You may not use the Service if you are located in a country subject to a comprehensive embargo by the United Nations, the European Union, the United States of America (including OFAC sanctions), or Mexico, or if you are on any applicable restricted party list.

1.6. The Service is currently not directed at residents or businesses located in the European Economic Area (EEA). Delph-AI has not yet designated an EU Representative under Article 27 of the GDPR. If your country of residence or primary place of business is in the EEA, you are not eligible to use the Service at this time. The Service may become available in the EEA in the future. If you have questions, contact legal@delph-ai.org.

2. Definitions

In these Terms, the following terms have the meanings set forth below:

• "Account" means your registered user account on the Platform.
• "Agreement Rate" means the level of consensus among AI Models on a given evaluation, expressed as a value between 0 and 1. The Agreement Rate is calculated using only valid, parseable model responses — responses that cannot be parsed into a valid include or exclude judgment are excluded from the calculation.
• "AI Models" means the third-party large language models and machine learning models used by the Service to evaluate Bibliographic Records, provided by Third-Party AI Providers.
• "Bibliographic Records" means the academic publication data (titles, abstracts, authors, and metadata) that you upload to the Service as part of a Dataset.
• "Dataset" means a collection of Bibliographic Records uploaded to the Service for use in a Screening.
• "Platform" means the Delph-AI web application, APIs, and related infrastructure.
• "Review Data" means all data you upload to, create within, or generate through the Service, including Datasets, Criteria, Screening configurations, and results.
• "Screening" means the automated evaluation process in which AI Models assess Bibliographic Records against user-defined inclusion and exclusion Criteria.
• "Criteria" means the inclusion and exclusion rules defined by you that AI Models use to evaluate Bibliographic Records.
• "Third-Party AI Providers" means the companies that provide AI Models or route requests to AI Models used by the Service, including but not limited to OpenAI, Anthropic, Google, Mistral, xAI, and DeepSeek, as well as models developed by Meta and Alibaba Cloud and accessed through Google Vertex AI.
• "Records" means the Bibliographic Records processed by the Service. Records are the unit of billing — each Record in your dataset counts toward your plan's allowance. Records are counted based on the gross number of entries in the Dataset at the time of submission, before deduplication or data-cleaning steps. A duplicate entry counts as a separate Record even if it is subsequently removed and only the unique entry is evaluated.
• "Version" means a recalculation of Screening results based on modified exclusion criteria, without re-triggering AI evaluations.
• "Merchant of Record" or "MoR" means a third party authorized to act as the legal seller or reseller for certain purchase transactions, handling checkout, payment collection, tax calculation, invoicing or receipts, fraud screening, payment-related refunds, and chargebacks. Delph-AI remains the supplier and provider of access to the Service unless expressly stated otherwise.
• "Buyer Terms" means the terms, policies, and other buyer-facing conditions of the applicable Merchant of Record that apply to a transaction processed through the Merchant of Record.
• "CFDI" means a Comprobante Fiscal Digital por Internet (Mexican electronic tax invoice) issued under applicable Mexican tax rules.

3. Description of Service

3.1. Delph-AI is a software-as-a-service platform that assists researchers in the title and abstract screening phase of systematic literature reviews. The Service uses multiple AI Models from different Third-Party AI Providers to evaluate Bibliographic Records against user-defined Criteria using a multi-model consensus method inspired by the Delphi method.

3.2. The Service is a tool to support the screening process. It does not replace human judgment, professional expertise, or the methodological rigor required for systematic reviews. For details on how our AI works, see our How Our AI Works page.

3.3. The Service is provided subject to availability. We may modify, update, suspend, or discontinue any part of the Service at any time. Material changes to the Service will be communicated in accordance with Section 18.

3.4. Features, pricing, and supported AI Models may change over time. We do not guarantee that any specific AI Model will remain available indefinitely.

4. User Accounts

4.1. To use the Service, you must create an Account by providing accurate, complete, and current information. You agree to keep your Account information updated.

4.2. You are responsible for maintaining the confidentiality of your Account credentials and for all activities that occur under your Account, whether or not authorized by you.

4.3. You must not share your Account credentials with any other person. Each Account is for a single individual user. If you need multiple users, you must obtain separate Accounts or a Team or Enterprise plan.

4.4. You must notify us immediately at security@delph-ai.org if you become aware of any unauthorized access to or use of your Account.

4.5. We reserve the right to suspend or terminate your Account if we reasonably believe that your Account has been compromised, that you have violated these Terms, or that your use of the Service poses a risk to us or other users.

5. Subscriptions, Records, and Payments

5.1. Plans and Pricing

The Service is available through the following subscription plans. A Free Trial of 500 records is available one time per payment method (card). Current prices for each plan are published on our Pricing page.

Pricing is subject to change with 30 days' notice. Changes to pricing for existing subscribers will take effect at the beginning of the next billing cycle following such notice.

Promotional and introductory pricing. Delph-AI may offer promotional or introductory subscription rates valid for a limited number of billing periods. The Early Access promotional rate applies to the first two (2) billing periods; beginning with the third billing period, the subscription automatically renews at the full recurring price disclosed at checkout. Other promotional offers may have different durations, which will be disclosed before purchase. Before purchase, the checkout page will identify the promotional price, the number of billing periods it applies to, and the recurring subscription price that will apply afterward. Your subscription will automatically renew at the stated recurring price after the promotional period ends, unless you cancel before the applicable renewal date.

5.2. Record Packages

You may purchase additional Record packages as one-time purchases. Record packages are available to any registered user, with or without an active subscription. They add records only — plan limits for criteria and models are not affected. Current record package prices and quantities are published on our Pricing page.

Model limit when using Record Packages without an active subscription: If you use purchased Record packages without an active paid subscription (for example, after cancelling your plan), you may submit Screenings using a maximum of 2 AI models and up to 5 criteria per Screening. This limit is automatically restored to your new plan's limit upon subscribing to or reactivating a paid plan.

5.3. Record Debit Policy

Records are debited at the time you submit a Screening for processing, not upon completion. The number of records debited equals the total number of Bibliographic Records in your dataset, based on the gross count at the time of submission before deduplication. Deduplication performed after submission does not reduce or reverse the debit; the number of Records billed may exceed the number of unique records ultimately evaluated. The costs of AI Model API calls are incurred at the time of processing, regardless of the outcome.

• If a Screening fails due to a technical error attributable to Delph-AI (such as a server crash or database failure), the debited records will be re-credited to your Account.
• If a Screening fails due to invalid input, insufficient data, or other causes attributable to you, records will not be re-credited.
• For the avoidance of doubt, inherent behaviors of Third-Party AI Models (including inaccurate outputs, hallucinations, unexpected classifications, or rate limiting by the AI provider) are not considered technical errors attributable to Delph-AI and do not give rise to record re-credits.

5.4. Record Balances and Expiration

Your account maintains two separate record balances:

• Subscription records (included in your plan) reset at the beginning of each billing period. Unused subscription records do not roll over to the next period.
• Purchased record packages do not expire and accumulate in a separate balance.
• When submitting a Screening, subscription records are consumed first (since they expire), followed by extra records.

5.5. Plan Limits

Each plan includes limits on the number of criteria, AI models, and reasoning models you may select per Screening. These limits cannot be exceeded by purchasing additional record packages — to increase these limits, you must upgrade your plan.

5.6. Service Limits

• You may have a maximum of 3 Screenings in "ongoing" (processing) status simultaneously, regardless of your plan. Attempting to submit a fourth will be rejected.
• Screenings in "ongoing" status cannot be cancelled by you. Only Delph-AI may cancel a Screening internally in exceptional circumstances.
• You may have a maximum of 100 active (non-deleted) projects at any time.

Screening processing times are not guaranteed. A Screening may be delayed due to Third-Party AI Provider availability, technical conditions, or retry operations. Contacting support regarding a delayed or stalled Screening does not create any right to cancellation, completion within a specific timeframe, or record re-credits. Cancellations and re-credits remain subject to Section 5.3. Nothing in this paragraph limits any right or remedy that cannot lawfully be excluded or limited under applicable law. Where a Screening stalls due to Third-Party AI Provider unavailability, Delph-AI may contact you to offer a resolution option, such as model substitution or a retry once the provider becomes available. If no resolution is accepted or mutually agreed upon, the Screening will remain in its stalled state. A failure to accept or agree on a resolution does not create any right to cancellation, refund, or record re-credits beyond those provided in Section 5.3.

When importing bibliographic records, abstracts exceeding 8,000 characters are trimmed to that limit before AI processing. This threshold is approximately 4–5 times the length of a standard academic abstract (250–350 words). Records exported with full-text content or clinical guideline text in the abstract field may be affected. You can view which records were trimmed in the dataset processing summary shown before screening submission.

5.7. Free Trial

The Free Trial provides 500 records as a one-time allocation. This allocation does not reset, replenish, or renew on a monthly or recurring basis. A valid payment method (credit or debit card) is required. Payment method verification is handled by our designated Merchant of Record (see Section 5.8.2). No charge is made during the trial. The Free Trial is limited to one per payment method (determined by card fingerprint). Abuse of the trial system may result in account restriction. Trial access may include only a subset of the AI Models available under paid subscription plans. The AI Models currently available to your Account are shown within the Service and are subject to change in accordance with Section 3.4.

5.8. Payments

All prices are expressed in United States Dollars (USD). Depending on your location and selected payment method, payments are processed through one of two channels described in Section 5.8.2. We never receive, store, or process your full credit card number or payment card details. All payment card data is handled by the applicable Merchant of Record in accordance with PCI-DSS standards.

5.8.1. Local Currency Pricing

When purchasing through a Merchant of Record, the checkout may display prices in your local currency. Currency availability, exchange rates, and the final amount charged are determined by the applicable Merchant of Record and displayed at checkout before you confirm payment. Your bank or card issuer may apply additional fees. Delph-AI does not provide currency exchange services or foreign exchange brokerage.

For Merchant of Record transactions, Delph-AI receives settlement payments in USD from the applicable Merchant of Record. Currency conversion is handled by the applicable Merchant of Record or your bank or card issuer. Exchange rates are determined by them and may differ from rates published by central banks. For approved SPEI transactions, the amount, currency, and applicable taxes will be stated in the invoice or payment instructions provided by Delph-AI. The amount displayed at checkout or stated in the applicable invoice shall prevail for the transaction.

Taxes, VAT, IVA, GST, sales tax, or similar charges may be calculated, collected, and remitted by the applicable Merchant of Record or payment channel, as shown at checkout. You are responsible for providing accurate billing, tax, and location information. Delph-AI is not responsible for errors caused by inaccurate information you provide.

5.8.2. Payment Channels

Delph-AI uses a dual-channel payment architecture:

(a) Merchant of Record Channel (International & Card Payments). When you purchase a subscription or record package using a credit or debit card, the transaction is processed by our designated Merchant of Record: currently Dodo Payments Inc. (primary) or Paddle (backup). The Merchant of Record is the legal seller or reseller for the payment transaction. It is responsible for checkout, payment collection, applicable taxes (such as VAT, GST, or sales tax), invoicing or receipts, fraud screening, payment-related refund processing, payment disputes, and chargebacks. By purchasing through a Merchant of Record, you also agree to their Buyer Terms presented during checkout. Delph-AI is the supplier and provider of access to the Service.

(b) SPEI Channel (Mexico Only). Mexican institutional customers that require a Mexican tax invoice (CFDI) may request to pay by direct SPEI bank transfer. In this channel, Delph-AI (Jesús Hernández Cabañas) is the direct seller and issues the applicable CFDI. The SPEI Channel is available only upon approval by Delph-AI and requires the customer to provide valid tax information (RFC). Applicable taxes (IVA) will be included as required by Mexican tax law.

Your applicable payment channel will be shown before purchase. If a conflict relates solely to checkout, payment collection, taxes, Merchant of Record-issued invoices or receipts, payment-related refund processing, chargebacks, or payment disputes in the Merchant of Record Channel, the applicable Merchant of Record's Buyer Terms will govern that payment matter, except where mandatory law provides otherwise. These Terms continue to govern your access to and use of the Service, including account access, acceptable use, support, and service availability. A payment channel's technical or commercial availability in a country does not mean that the Service is offered or available in that country.

5.9. Billing and Renewal

Subscriptions renew automatically at the end of each monthly billing period unless you cancel before the end of the current period. You may cancel your subscription at any time through your Account settings. Upon cancellation, you will retain access to the Service until the end of your current paid period.

If you request a downgrade to a lower-tier plan during a paid billing period, the downgrade takes effect at the beginning of your next billing period. Your current plan remains active until then. Downgrades do not result in prorated refunds or credits, except where required by applicable law. The limits and features of your new plan will apply from the effective date of the downgrade.

5.10. Refund Policy

Subscription fees and record package purchases are generally non-refundable, except:

(a) Where required by applicable law;

(b) If Delph-AI terminates your Account without cause, in which case we will refund the pro-rata portion of any prepaid subscription fees for the remaining period;

(c) If the Service is permanently discontinued, in which case we will refund unused prepaid subscription fees on a pro-rata basis.

Additional refund rights may apply to consumers in certain jurisdictions under mandatory applicable law.

6. Intellectual Property

6.1. The Service, including its source code, algorithms, user interface, design, branding, trademarks, and all related intellectual property, is and remains the exclusive property of Delph-AI and its licensors.

6.2. Subject to your compliance with these Terms, we grant you a limited, non-exclusive, non-transferable, non-sublicensable, revocable license to access and use the Service solely for its intended purpose.

6.3. Nothing in these Terms grants you any right to the source code, underlying algorithms, AI Models, or training data used by the Service or by Third-Party AI Providers.

6.4. All rights not expressly granted in these Terms are reserved by Delph-AI.

7. User Content

7.1. You retain full ownership of your Review Data. Delph-AI does not claim any ownership rights over your Datasets, Criteria, Screening configurations, results, or any other data you upload to or create within the Service.

7.2. You grant Delph-AI a limited, non-exclusive license to use, process, store, and transmit your Review Data solely for the purpose of operating, maintaining, and providing the Service to you. This license terminates when you delete your data or your Account, subject to Section 17 (Data Retention and Deletion).

7.3. Delph-AI does not use your Review Data to train, fine-tune, or improve any AI models. Our agreements with Third-Party AI Providers include terms that prohibit the use of API data for model training.

7.4. You represent and warrant that you have the right to upload and use all Review Data that you submit to the Service, and that your Review Data does not infringe any third party's intellectual property rights or violate any applicable law. If your Bibliographic Records originate from third-party databases (such as Scopus, Embase, Web of Science, PubMed, or similar), you are responsible for complying with the applicable license terms of those databases.

7.5. You are solely responsible for the accuracy, quality, legality, and appropriateness of your Review Data. Delph-AI does not verify the accuracy or completeness of Bibliographic Records you upload.

8. Aggregated and Anonymized Data

8.1. Delph-AI may create, use, and retain aggregated and anonymized data derived from your use of the Service. Aggregated and anonymized data does not identify you or any individual user.

8.2. We may use aggregated and anonymized data for:

• Improving and developing the Service;
• Internal analytics and benchmarking;
• Research and statistical purposes;
• Generating industry insights (e.g., average screening sizes, model performance trends).

8.3. We will never sell aggregated or anonymized data to third parties, use it for targeted advertising, or use it to train AI models in a way that could reconstruct individual user data.

9. AI-Specific Terms

This section describes important terms specific to the AI-powered features of the Service. Please read it carefully.

9.1. Nature of AI Outputs

The Service uses multiple AI Models from Third-Party AI Providers to evaluate Bibliographic Records. The evaluations, classifications, Agreement Rates, and any other outputs generated by the AI Models ("AI Outputs") are produced by probabilistic machine learning systems and:

(a) May contain errors, inaccuracies, or omissions. AI Models may produce incorrect evaluations, including false inclusions, false exclusions, or inconsistent results;

(b) Are not verified for accuracy by Delph-AI. We do not manually review individual AI evaluations;

(c) May vary between executions. Running the same Screening with the same parameters may produce slightly different results due to the probabilistic nature of AI Models;

(d) May be similar to outputs provided to other users. If other users evaluate similar Bibliographic Records with similar Criteria, they may obtain similar or identical results. This is expected behavior, not a defect.

9.2. AI Outputs Are Not Professional Advice

AI Outputs are provided as decision-support tools and do not constitute medical, scientific, legal, or professional advice of any kind. Specifically:

(a) AI Outputs do not replace human review of a systematic literature review. Current best practices (including PRISMA, the Cochrane Handbook, and the RAISE reporting guidelines) recommend human oversight of AI-assisted screening;

(b) AI Outputs must not be used as the sole basis for clinical decisions, treatment recommendations, public health policies, or any other decision that may affect human health or safety;

(c) The use of Delph-AI does not guarantee that your systematic review will comply with the standards of any organization, including Cochrane, the Joanna Briggs Institute (JBI), the Campbell Collaboration, or any journal's publication requirements;

(d) The quality of AI Outputs depends on the quality of your inputs. The accuracy of Screening results is directly affected by the clarity and specificity of your Criteria, the completeness of your Bibliographic Records, and the appropriateness of the AI Models you select.

9.3. Your Obligation to Verify

You agree to independently review and verify all AI Outputs before relying on them for inclusion/exclusion decisions in your systematic review or for any other purpose. You accept full responsibility for how you use AI Outputs.

9.4. Multi-Model Consensus

The Service uses a consensus mechanism in which multiple AI Models independently evaluate each Bibliographic Record. The Agreement Rate reflects the level of concordance among valid parsed evaluations — responses that cannot be parsed into a valid include or exclude judgment are excluded from the calculation. A higher Agreement Rate generally indicates greater confidence in the evaluation, but does not guarantee correctness. The final classification is determined by a weighted consensus, not by any single model.

9.5. No Training with Your Data

Delph-AI does not use your Review Data, Criteria, or AI Outputs to train, fine-tune, or improve any machine learning models. Our agreements with Third-Party AI Providers include terms that prohibit the use of API inputs and outputs for model training.

9.6. Third-Party AI Models

The AI Models used by the Service are provided by third-party companies (currently including OpenAI, Anthropic, Google, Mistral, xAI, and DeepSeek, as well as models developed by Meta and Alibaba Cloud and accessed through Google Vertex AI). Delph-AI does not own, control, or guarantee the availability, accuracy, performance, or continued existence of any Third-Party AI Model. See Section 12 for more information.

10. Prohibited Uses

You agree not to use the Service for any of the following purposes. Violation of this section may result in immediate suspension or termination of your Account without prior notice.

10.1. Prohibited Data

You must not upload, submit, or process through the Service:

(a) Personal health information (PHI) or any data protected under HIPAA, or special categories of personal data as defined in Article 9 of the GDPR (including health data, genetic data, or biometric data);

(b) Personal data of minors under the age of 16;

(c) Classified, military, or national security information;

(d) Personal financial data (credit card numbers, bank account numbers, social security numbers).

10.2. Prohibited Uses of the Service

You must not:

(a) Use AI Outputs as the sole basis for clinical decisions, medical treatment, or public health policies without qualified professional supervision;

(b) Use the Service for military purposes, surveillance, weapons development, or intelligence gathering;

(c) Redistribute, resell, sublicense, or make the Service available to third parties, whether for a fee or free of charge;

(d) Use the Service to build a competing product or service, or copy any features, functions, or graphics of the Service;

(e) Use AI Outputs to train, fine-tune, or improve any machine learning or artificial intelligence model, whether your own or a third party's;

(f) Provide the Service to third parties on a time-sharing, service bureau, or similar basis;

(g) Engage in automated scraping, crawling, or mass data extraction from the Service.

10.3. Prohibited Conduct

You must not:

(a) Reverse engineer, decompile, disassemble, or attempt to derive the source code of the Service;

(b) Circumvent, disable, or interfere with any usage limits, security measures, or access controls of the Service;

(c) Interfere with the integrity, security, or availability of the Service or the experience of other users;

(d) Share your Account credentials or provide access to non-public features to unauthorized persons;

(e) Misrepresent AI Outputs as human-authored reviews or as endorsed or approved by Delph-AI;

(f) Infringe the intellectual property rights of any third party through your use of the Service;

(g) Violate any applicable law, regulation, or third-party right through your use of the Service.

11. Disclaimer of Warranties

EXCEPT FOR THE EXPRESS COMMITMENTS SET OUT IN THESE TERMS, AND WITHOUT LIMITING ANY NON-WAIVABLE CONSUMER, PRIVACY, DATA PROTECTION, OR OTHER STATUTORY RIGHTS, THE SERVICE IS PROVIDED ON AN "AS IS" AND "AS AVAILABLE" BASIS. TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, DELPH-AI DISCLAIMS ANY WARRANTIES, CONDITIONS, OR REPRESENTATIONS NOT EXPRESSLY STATED IN THESE TERMS, INCLUDING WARRANTIES OF ACCURACY, RELIABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE, NON-INFRINGEMENT, AND UNINTERRUPTED OR ERROR-FREE OPERATION.

DELPH-AI DOES NOT WARRANT THAT:

(A) THE SERVICE WILL BE UNINTERRUPTED, TIMELY, SECURE, OR ERROR-FREE;

(B) THE RESULTS OBTAINED FROM THE USE OF THE SERVICE, INCLUDING AI OUTPUTS, WILL BE ACCURATE, RELIABLE, OR COMPLETE;

(C) THE SERVICE WILL MEET YOUR REQUIREMENTS OR EXPECTATIONS;

(D) ANY ERRORS IN THE SERVICE WILL BE CORRECTED;

(E) THE SERVICE WILL BE COMPATIBLE WITH ANY PARTICULAR HARDWARE, SOFTWARE, OR NETWORK CONFIGURATION.

YOU ACKNOWLEDGE THAT AI OUTPUTS ARE GENERATED BY PROBABILISTIC MACHINE LEARNING MODELS AND THAT SUCH OUTPUTS MAY CONTAIN ERRORS, HALLUCINATIONS, OR INACCURACIES. YOU USE THE SERVICE AND RELY ON AI OUTPUTS ENTIRELY AT YOUR OWN RISK.

Some jurisdictions do not allow the exclusion of certain warranties. In such jurisdictions, the above exclusions apply to the maximum extent permitted by applicable law.

12. Third-Party Services and AI Providers

12.1. The Service relies on third-party services and AI Models to function, including but not limited to:

• AI Model Providers: OpenAI, Anthropic, Google (Vertex AI), Mistral, xAI, DeepSeek (direct API), Meta (model developer; via Vertex AI), Alibaba Cloud (model developer; via Vertex AI)
• Cloud Infrastructure: Google Cloud Platform
• Authentication: Firebase (Google)
• Merchant of Record / Payment Processing: Dodo Payments Inc. (primary), Paddle (backup)
• Transactional Email: Resend

For a complete and current list of service providers, see our Service Providers & Sub-Processors page.

12.2. Delph-AI is not responsible for the acts, omissions, availability, performance, or policies of any third-party service provider. Third-party services are subject to their own terms of service and privacy policies.

12.3. Delph-AI is not liable for any intellectual property claims arising from the underlying AI Models or their training data, which are developed, owned, and operated by Third-Party AI Providers.

12.4. Third-Party AI Providers may change, deprecate, or discontinue their models or services at any time. Delph-AI will make commercially reasonable efforts to maintain the availability of the Service, but does not guarantee the continued availability of any specific AI Model.

13. Limitation of Liability

13.1. TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, THE TOTAL AGGREGATE LIABILITY OF DELPH-AI ARISING OUT OF OR RELATED TO THESE TERMS OR THE SERVICE SHALL NOT EXCEED THE GREATER OF: (A) ONE HUNDRED U.S. DOLLARS ($100 USD); OR (B) THE TOTAL FEES PAID BY YOU TO DELPH-AI DURING THE TWELVE (12) MONTHS IMMEDIATELY PRECEDING THE EVENT GIVING RISE TO THE CLAIM.

13.2. EXCEPT FOR LIABILITY THAT CANNOT BE LIMITED UNDER APPLICABLE LAW AND SUBJECT TO SECTION 13.3, DELPH-AI SHALL NOT BE LIABLE FOR INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, EXEMPLARY, PUNITIVE, OR NON-IMMEDIATE DAMAGES, INCLUDING BUT NOT LIMITED TO DAMAGES FOR:

(A) LOSS OF PROFITS, REVENUE, OR BUSINESS;

(B) LOSS OF DATA OR REVIEW DATA;

(C) LOSS OF GOODWILL OR REPUTATION;

(D) BUSINESS INTERRUPTION;

(E) COST OF PROCUREMENT OF SUBSTITUTE SERVICES;

(F) ANY DECISIONS MADE OR ACTIONS TAKEN IN RELIANCE ON AI OUTPUTS;

WHETHER BASED ON WARRANTY, CONTRACT, TORT (INCLUDING NEGLIGENCE), STRICT LIABILITY, OR ANY OTHER LEGAL THEORY, EVEN IF DELPH-AI HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

13.3. The limitations in this section do not apply to: (a) Delph-AI's willful misconduct or gross negligence; (b) fees owed by you to Delph-AI. For the avoidance of doubt, indemnification obligations under Section 14 are subject to the aggregate liability cap in Section 13.1.

13.4. Some jurisdictions do not allow the exclusion or limitation of certain damages. In such jurisdictions, the liability of Delph-AI shall be limited to the maximum extent permitted by applicable law.

14. Indemnification

14.1. Your Indemnification of Delph-AI

You agree to indemnify, defend, and hold harmless Delph-AI, its officers, directors, employees, and agents from and against any and all claims, damages, losses, liabilities, costs, and expenses (including reasonable attorneys' fees) arising out of or related to:

(a) Your Review Data, including any claim that your Review Data infringes any third party's rights;

(b) Your violation of Section 10 (Prohibited Uses);

(c) Your use of AI Outputs, including any decisions made or actions taken based on AI Outputs, except to the extent directly caused by a defect in the Service attributable to Delph-AI;

(d) Your violation of any applicable law or regulation;

(e) Your breach of these Terms.

14.2. Delph-AI's Indemnification of You

Delph-AI will indemnify, defend, and hold you harmless from and against any third-party claim alleging that the Service itself (excluding your Review Data, AI Outputs, and Third-Party AI Models) infringes any third party's intellectual property rights, provided that you: (a) promptly notify Delph-AI of the claim; (b) grant Delph-AI sole control of the defense and settlement; and (c) cooperate with Delph-AI as reasonably requested.

14.3. Exclusions from Delph-AI's Indemnification

Delph-AI shall not be obligated to indemnify you for claims arising from: (a) AI Outputs or their use; (b) the intellectual property of AI Models, training data, or other materials provided by Third-Party AI Providers; (c) your Review Data; (d) modifications you make to the Service or its outputs; (e) your use of the Service in violation of these Terms.

15. Force Majeure

15.1. Neither party shall be liable for any delay or failure to perform its obligations under these Terms (other than payment obligations) if such delay or failure results from circumstances beyond its reasonable control, including but not limited to:

(a) Natural disasters, epidemics, pandemics, fires, floods, or severe weather;

(b) War, terrorism, civil unrest, sanctions, or embargoes;

(c) Government actions, regulations, or orders;

(d) Failures, outages, or service disruptions of Third-Party AI Providers (including changes to their APIs, pricing, or availability);

(e) Failures or outages of cloud infrastructure providers (including Google Cloud Platform);

(f) Cyberattacks, distributed denial-of-service (DDoS) attacks, or other security incidents beyond the affected party's reasonable control;

(g) Labor disputes, strikes, or shortages;

(h) Internet or telecommunications failures.

15.2. The affected party shall notify the other party without undue delay and shall use commercially reasonable efforts to mitigate the impact of the force majeure event.

15.3. If a force majeure event continues for more than 90 consecutive days, either party may terminate these Terms upon written notice to the other party.

16. Termination and Cancellation

16.1. Cancellation by You

You may cancel your subscription at any time through your Account settings. Upon cancellation:

(a) You will retain access to the Service until the end of your current paid billing period;

(b) No further charges will be incurred after the current period ends;

(c) Unused subscription records for the current period will remain available until the period ends, after which they reset;

(d) Any purchased Record packages (extra records) you have accumulated will be retained indefinitely in your account. However, once your subscription ends, you may only use those records with a maximum of 2 AI models and 5 criteria per Screening until you subscribe to a paid plan again.

16.2. Termination by Delph-AI for Cause

Delph-AI may terminate or suspend your Account:

(a) With 15 days' written notice if you breach these Terms and the breach is curable, to give you an opportunity to cure;

(b) Immediately and without notice if: (i) you engage in fraud; (ii) you violate Section 10 (Prohibited Uses); (iii) your continued use poses an imminent risk to other users, the Service, or Delph-AI; or (iv) we are required to do so by law.

16.3. Termination by Delph-AI Without Cause

Delph-AI may terminate your Account without cause upon 30 days' written notice. In such case, we will refund the pro-rata portion of any prepaid subscription fees for the remaining period.

16.4. Discontinuation of the Service

Delph-AI may discontinue the Service entirely upon 90 days' written notice to all users. In such case, we will refund unused prepaid subscription fees on a pro-rata basis.

16.5. Effect of Termination

Upon termination:

(a) Your license to use the Service immediately terminates;

(b) You may export your Review Data during the applicable grace period (see Section 17);

(c) Sections 6, 7.3, 8, 9, 11, 13, 14, 17, and 19 survive termination.

17. Data Retention and Deletion

17.1. Upon deletion of your Account, the following grace periods apply:

17.2. During the applicable grace period, you may restore your Account, projects, or screenings. After the grace period expires, anonymization is irreversible.

17.3. For complete information about how we handle your data, please see our Privacy Policy.

18. Modifications and Notifications

18.1. Delph-AI reserves the right to modify these Terms at any time. For material changes:

(a) We will provide at least 30 days' notice before the changes take effect;

(b) Notice will be sent to the email address associated with your Account and displayed as a prominent banner within the Service;

(c) The notice will include a summary of the material changes and a link to the updated Terms;

(d) If you do not agree with the changes, you may cancel your Account before the changes take effect and receive a pro-rata refund of prepaid subscription fees for the remaining period.

18.2. Your continued use of the Service after the 30-day notice period constitutes acceptance of the modified Terms.

18.3. Non-material changes (such as typographical corrections or clarifications that do not affect your rights or obligations) may take effect immediately upon posting.

18.4. Changes required by law, regulation, or court order may take effect immediately without the 30-day notice period.

18.5. We will make previous versions of these Terms available upon request where reasonably practicable.

19. Governing Law, Dispute Resolution, and General Provisions

19.1. Governing Law

These Terms shall be governed by and construed in accordance with the laws of the United Mexican States (Mexico), without regard to conflict of law principles.

19.2. Mandatory Consumer Protections

Nothing in these Terms limits or excludes any non-waivable consumer, privacy, or other statutory rights that apply to your lawful use of the Service under mandatory law. References to mandatory laws in any jurisdiction apply only to the extent those laws apply to your use of the Service and do not create eligibility to access the Service from a jurisdiction where access is restricted under Section 1.6.

19.3. Dispute Resolution

Any dispute arising out of or in connection with these Terms shall be resolved as follows:

(a) Negotiation (30 days). The parties shall first attempt to resolve the dispute through good-faith negotiation. Either party may initiate this process by sending a written notice describing the dispute to the other party.

(b) Mediation (30 days). If negotiation fails, the parties shall attempt mediation before a mutually agreed mediator.

(c) Litigation. If mediation fails, the dispute shall be submitted to the exclusive jurisdiction of the competent courts of Mexico City, Mexico, except where mandatory consumer protection laws grant you the right to bring proceedings in your local courts.

(d) Small Claims Exception. Notwithstanding the above, either party may bring an individual action in a court of competent jurisdiction for disputes involving amounts not exceeding $10,000 USD.

19.4. No U.S.-Style Class Actions

To the maximum extent permitted by applicable law, the parties agree that disputes will be resolved only between the parties and not as a U.S.-style class, consolidated, or representative action. Nothing in this Section limits or waives any non-waivable right to participate in or bring collective actions or representative proceedings available under Mexican consumer protection or procedural law.

19.6. Severability

If any provision of these Terms is held to be invalid, illegal, or unenforceable by a court of competent jurisdiction, the remaining provisions shall continue in full force and effect. The invalid provision shall be modified to the minimum extent necessary to make it valid and enforceable while preserving its original intent.

19.7. Assignment

You may not assign or transfer these Terms or your rights under them without our prior written consent. Delph-AI may assign these Terms in connection with a merger, acquisition, or sale of all or substantially all of its assets, provided that the assignee agrees to be bound by these Terms.

19.8. Waiver

The failure of either party to enforce any provision of these Terms shall not constitute a waiver of that provision or the right to enforce it in the future.

19.9. Entire Agreement

These Terms, together with the Privacy Policy, the Service Providers & Sub-Processors page, and the How Our AI Works page, constitute the entire agreement between you and Delph-AI with respect to the Service and supersede all prior agreements and understandings.

19.10. Contact Us

If you have any questions about these Terms, please contact us:

• Legal inquiries: legal@delph-ai.org
• Privacy inquiries: privacy@delph-ai.org
• Security reports: security@delph-ai.org
• General support: support@delph-ai.org