AI Use and Transparency.

Effective date: 2026-10-27.

This AI Use and Transparency Policy (this "Policy") explains how wrxstack ("wrxstack", "we", "us") uses artificial intelligence ("AI") in its products, how data moves when you use an AI feature, and the rules that apply to AI output. It applies to the wrxstack products, including Atlas at atlas.wrxstack.com and Folio, the portfolio and personal-site builder, at portfolio.wrxstack.com (together, the "Service").

This Policy should be read with our Terms of Service, Acceptable Use Policy, and Privacy Policy. Terms with capital letters that are not defined here, such as "Customer Content" and "AI Output", have the meanings given in the Terms of Service. If this Policy and the Terms of Service conflict, the Terms of Service control. Section 22 explains the legal effect of this Policy.

The short version

1. Scope and roles

This Policy covers every feature of the Service that sends information to an AI model to produce a result, which we call an "AI feature". It covers the information you give to an AI feature (your "Input") and the result the AI feature produces (the "AI Output").

Our role. wrxstack builds the AI features and decides how they work. We do not develop or train general-purpose AI models of our own. The AI features are built on models made available by third-party model providers, as described in section 4.

Your role. When you use an AI feature, you decide what Input to give it and what to do with the AI Output. Where you use the Service through a workspace that your employer or another organization controls (the "Customer"), that Customer decides how AI features are used in its workspace, is the controller of the personal data in it, and is responsible for its own legal duties as a user of AI. We process that workspace data on the Customer's instructions under our Data Processing Addendum.

Who operates wrxstack. wrxstack is an independent, privately owned product run by a small team. Questions about this Policy go to contact@wrxstack.com.

2. Our principles for AI

We design and operate the AI features of the Service on the following principles. Each is described in more detail later in this Policy.

• Your content stays yours. We do not use your Input or AI Output to train any model, and we select model providers whose API terms prohibit them from doing so.
• Use on request. The Service is designed to send content to a model only to carry out an AI feature that a person has chosen to use.
• Least data. We aim to design each AI feature to send the model only the information it needs to carry out the request.
• Openness. We aim to make it clear when you are using AI and which kinds of providers process your content.
• People decide. AI Output supports human judgment. It does not replace it, and the Service does not use AI to make decisions about people that have legal or similarly significant effects.
• Honesty about limits. We do not claim that AI Output is accurate, complete, or free from bias.

3. Where the Service uses AI

The table below describes the main AI features of each product. Products change over time, and a product may offer additional AI features. Any AI feature, whether or not it is listed here, is covered by this Policy.

The Service does not use AI for advertising, for selling data, for scoring or ranking people, or for emotion recognition, biometric identification, or biometric categorization.

4. Models and model providers

4.1 Third-party model providers

Unless a workspace has chosen its own model under section 4.2, AI features are carried out by third-party model providers that we select. We access their models through their commercial application programming interfaces (APIs). We choose providers whose API terms prohibit the use of submitted content to train models, and we may use more than one provider, or change providers, to keep the Service working well. Where a product lets you pick a model from a list, the provider of that model is the one that processes your request.

Our sub-processor list names the service providers that process Customer Personal Data for us, and we aim to keep it up to date as that page describes. A model provider receives only the content sent to it for an AI feature, at the time the feature is used.

4.2 Bring your own model

Where a product offers a "bring your own model" option and a workspace uses it, the workspace connects a model provider or endpoint that it chooses, using its own account and credentials. In that case:

• the content for those AI features is sent to the provider the workspace chose, not to a provider chosen by us;
• the relationship between the workspace and that provider, including how the provider stores, uses, or trains on content, is governed by the workspace's own agreement with that provider;
• that provider is not our sub-processor, and we are not responsible for its acts, its terms, or its availability;
• the workspace is responsible for checking that the provider's terms meet its own legal, contractual, and confidentiality needs.

5. What is sent to a model, and when

5.1 When

The Service is designed so that content is sent to a model only when a person uses an AI feature, for example by typing a request to the assistant, selecting a generate or suggest button, or running an AI step in a workflow that a person in the workspace has set up. We do not run background jobs that send your content to a model provider on our own initiative. If no one in your workspace uses an AI feature, the Service does not send content from it to a model provider.

5.2 What

For each request, the AI feature sends the model:

• the Input you give it, such as your question, instruction, or pasted text;
• any file or text you choose to attach or select for the request, such as the text extracted from a résumé PDF;
• the context the feature needs to carry out the request, such as the title and text of the page for which you asked for a search engine description, or the content of the workspace items the request relates to;
• instructions written by wrxstack that tell the model how to carry out the feature.

AI features are designed not to send your password, sign-in tokens, API keys, or payment details to a model provider. Where a request is carried out inside a workspace, the AI feature works within the access that the person making the request already has.

5.3 Sensitive information

Do not give an AI feature information that you are not permitted to share, or special categories of personal data, such as health, biometric, or financial account information, unless you have a lawful basis to process it and it is necessary for your purpose. Anything you include in Input is sent to the model provider for that request.

6. No training on your content

We do not use your Input, your Customer Content, or AI Output to train, fine-tune, or evaluate any AI model. We do not share them with model providers for those purposes.

Model providers we select receive content under their commercial API terms, which prohibit them from using that content to train their models. This prohibition is a term of their agreements with us. We do not control how a provider operates, and we are not responsible for a provider's failure to follow its own terms. Where a workspace uses its own model under section 4.2, the training terms are those of the workspace's own agreement with its chosen provider.

7. Retention of AI data

We design the Service so that the text of your AI requests is not written to our application logs.

8. Transparency and disclosure

8.1 What we tell you

Consistent with the transparency duties of the EU Artificial Intelligence Act (Regulation (EU) 2024/1689), including Article 50, and similar laws elsewhere, we take reasonable steps to:

• present AI features so that it is clear you are using AI, and not present an AI feature as a human being;
• answer truthfully if you ask whether you are interacting with AI;
• publish this Policy, which explains what AI features do, what data they use, and their limits;
• identify, in section 4 and our sub-processor list, the kinds of providers that process your content.

8.2 What you must tell others

If you use AI Output with other people, you are responsible for any disclosure the law requires of you. In particular:

• if you use AI Output to create images, audio, or video that look like real people, places, or events and could falsely appear authentic, you must disclose that the content was generated or altered by AI;
• if you publish AI-generated text to inform the public on matters of public interest, you must disclose that it was generated by AI, unless a person has reviewed it and taken editorial responsibility for it where the law allows that exception;
• if you use the Service to build a chatbot, agent, or other system that talks to other people, you must make sure those people know they are interacting with AI;
• you must not remove or alter any label, notice, or provenance information that identifies content as AI-generated, for the purpose of avoiding a disclosure duty.

These duties are set out in our Acceptable Use Policy and apply in addition to this Policy.

9. Human oversight and control

• You choose when AI runs. AI features run because a person chose to use them, as described in section 5.1.
• You review before you rely. AI Output is shown to you as a draft or suggestion. You can edit, accept, or discard it.
• Actions follow your instructions. Where an AI feature can take an action, such as creating or changing an item or sending data to a connected tool, it does so because a person asked it to or set it up to, and within that person's permissions. Workspace administrators can control access through role-based access, and actions in Atlas are recorded in its audit log.
• You can stop using AI. You can use the Service without using any AI feature. Not using AI features is the simplest way to make sure none of your content is sent to a model provider.

10. Accuracy and limits of AI

AI models produce output by predicting likely text from patterns. They do not understand meaning, check facts, or know the truth in the way a person does. As a result, AI Output:

• can be wrong, incomplete, or out of date, and can state false information confidently;
• can invent facts, quotations, sources, figures, or references that do not exist;
• can reflect bias found in the data on which the model was trained;
• can differ each time, even for the same Input;
• can be similar or identical to output produced for other users;
• can be affected by instructions hidden in content you ask it to process, as described in section 17.

AI Output is not professional advice. It is not legal, medical, financial, tax, investment, employment, or safety advice, and it must not be treated as a substitute for advice from a qualified professional. We make no promise that AI Output is accurate, complete, original, or fit for any purpose. The disclaimers in the Terms of Service apply to AI Output.

11. Your responsibility for AI Output

You are responsible for your Input and for what you do with AI Output. In particular, you must:

• have the right to use any content you give an AI feature, including the right to share any personal data in it;
• review AI Output, and check facts, figures, names, and references, before you rely on it, publish it, or act on it;
• not rely on AI Output alone for any decision that matters, including any decision about a person;
• make sure your use of AI Output complies with the law, with the rights of others, and with the Acceptable Use Policy;
• make any disclosure the law requires, as described in section 8.2.

For a portfolio, résumé, or similar profile, check that everything the AI feature produced about your experience, qualifications, and achievements is true before you publish it. You are responsible for the accuracy of what you publish about yourself.

12. No automated decisions about people

wrxstack does not use AI, or any other automated processing, to make decisions about you that produce legal effects or similarly significant effects, within the meaning of Article 22 of the EU and UK General Data Protection Regulation. We do not use AI to decide whether you may open or keep an account, to set your price, or to profile you for such decisions.

We may use automated tools to protect the Service, such as rate limits and abuse checks. Where an automated check leads to a restriction on your account, you can ask us to review it by emailing contact@wrxstack.com, and we will consider your request.

A Customer that uses AI Output to make decisions about its own staff, applicants, customers, or other people is responsible for those decisions, including any duty to provide human review, an explanation, or a right to contest under the law that applies to it. See section 14.

13. Prohibited AI uses

The Acceptable Use Policy sets out the uses of the Service, including its AI features, that are not allowed. Without limiting it, you must not use any AI feature to:

• carry out any practice prohibited by Article 5 of the EU Artificial Intelligence Act, including manipulating people through subliminal or deceptive techniques that cause harm, exploiting people's vulnerabilities due to age, disability, or social or economic situation, social scoring, predicting the risk that a person will commit a crime based only on profiling, building facial recognition databases by untargeted scraping, inferring emotions in the workplace or in education, or categorizing people by biometric data to infer sensitive characteristics;
• impersonate a real person or organization, or create deceptive content, disinformation, or fabricated evidence;
• create non-consensual intimate imagery, sexual content involving minors, or content that harasses, threatens, or promotes violence or hatred;
• create malware, plan attacks on systems, or obtain help with weapons capable of mass harm;
• process personal data unlawfully, including by collecting or inferring information about people without a lawful basis;
• bypass the safety measures of the Service or of a model, or try to extract system instructions, model weights, or training data;
• breach the usage policies of the model provider that processes the request.

We may refuse, limit, or remove access to AI features, or take the other steps set out in the Acceptable Use Policy, where we reasonably believe a use breaks these rules.

14. High-risk and regulated uses

The Service is a general-purpose workspace and portfolio tool. It is not designed, tested, or intended for use as a system that decides or materially influences outcomes for people in areas the law treats as high risk, such as:

• recruitment, hiring, promotion, termination, or monitoring of workers;
• access to education, or assessment of students;
• credit, lending, insurance, or housing;
• access to essential public or private services and benefits;
• healthcare, diagnosis, or treatment;
• law enforcement, migration, border control, or the administration of justice;
• the safety components of critical infrastructure.

If you choose to use the Service in any of these areas, you do so at your own decision and are solely responsible for meeting every legal requirement that applies to that use, which may include requirements under the EU Artificial Intelligence Act, data protection law, the California rules on automated decision-making technology, and US state and local laws on AI in employment and consumer decisions. That includes carrying out any required assessment, giving any required notice, and making sure a qualified person reviews each decision.

15. Intellectual property

• Your Input. You keep all rights in your Input. You give us only the permissions we need to provide the Service, as set out in the Terms of Service.
• AI Output. As between you and wrxstack, you own the AI Output generated from your Input, as the Terms of Service provide. wrxstack claims no ownership of it.
• Limits on ownership. The law in many places, including the United States, may not protect material that was created by AI without enough human creative input. Whether AI Output can be protected by copyright depends on the law where you are and on your own contribution to it. Because AI Output can be similar or identical to output produced for others, you may not have exclusive rights in it.
• Rights of others. We do not promise that AI Output is free from the rights of others. Before you use AI Output commercially or publish it, check that it does not copy protected material or use another person's name, likeness, or trademark without permission.
• Copyright complaints. If you believe content on the Service infringes your rights, follow our Copyright Policy.

16. Integrations, connected tools, and actions

Atlas can work with integrations, webhooks, its REST API, and tools connected through the Model Context Protocol (MCP). When you connect a tool and ask an AI feature to use it:

• the AI feature may read from or send data to that tool, as your request requires;
• data you send to a third-party tool is handled under that third party's terms and privacy policy, not ours;
• you are responsible for deciding which tools to connect, which permissions to grant them, and which actions to allow;
• you should connect only tools you trust, and review actions that change or send data before you confirm them where the product asks you to.

We are not responsible for the acts of third-party tools, or for actions that a person in your workspace instructed an AI feature to take.

17. Security of AI features

AI features are protected by the same measures as the rest of the Service: encryption in transit using TLS 1.3, encryption at rest provided by our hosting provider, and access controls. These measures are described on our Security page and in Data handling. wrxstack holds no security certification and has not had a third-party audit.

AI features carry a risk known as prompt injection: a document, web page, email, or other content that an AI feature processes may contain hidden instructions that try to change what the feature does. We work to limit this risk, but no current method removes it completely. To reduce it, process content only from sources you trust, check AI Output before you act on it, and review any action an AI feature proposes before you confirm it.

If you find a security weakness in an AI feature, report it under our Responsible Disclosure Policy.

18. Your privacy rights and AI

The personal data in your Input and AI Output is covered by our Privacy Policy. You have the same rights over it as over any other personal data we hold, including the rights to access, correct, delete, and object, and the right not to be subject to decisions based solely on automated processing, as described in our GDPR Statement and California Notice.

To exercise a right, email contact@wrxstack.com. We respond within the period the law that applies to your request requires, which under the GDPR is one month, extendable by two further months for complex or numerous requests. California residents should see the California Notice for the timelines that apply to them. Where the data belongs to a Customer's workspace, we may pass the request to the Customer and assist it as the Data Processing Addendum and the law require. Once a model provider has returned a result, we cannot recall the content from it, but its handling of that content is limited by its API terms as described in section 7.

19. Children

The Service, including its AI features, is intended for people aged 16 and over. AI features must not be used by anyone younger, or to create content directed at or about children in a way that could harm them. If you believe a child under 16 has used the Service, email contact@wrxstack.com and we will take reasonable steps to delete their data, as the law requires.

20. Reporting a problem with AI

Please tell us if an AI feature produces output that is harmful, unlawful, seriously inaccurate, biased, or infringing, or if you believe an AI feature has been misused. Email contact@wrxstack.com with the subject "AI concern" and include:

• the product and feature concerned;
• the date and approximate time;
• the Input you used and the AI Output you received, or a description of them;
• why you believe the output or use is a problem.

Do not include passwords or payment details. We aim to review reports, and we may take the steps we consider appropriate, at our discretion, such as changing a feature's instructions, limiting a use, or acting under the Acceptable Use Policy. We tell you the outcome where the law requires it, and otherwise we may reply where you have asked for a reply. If the matter involves an imminent risk to someone's life or safety, contact your local emergency services first.

21. No compensation

This Policy explains how AI features work and the rules for using them. It does not create any right to a payment, credit, reimbursement, reward, or compensation of any kind, for you or for us, including in connection with any AI Output, any error in AI Output, or any report made under section 20. It does not change the disclaimers, the limitation of liability, or any other provision of the Terms of Service.

22. Legal effect of this Policy

This Policy explains our practices. It is not a contract, and it does not create any contractual obligation for wrxstack, or any warranty, guarantee, or right for you or for any other person. The only agreements between you and wrxstack are the Terms of Service and, where it applies, the Data Processing Addendum. The rules this Policy sets for your use of AI features, including sections 8.2, 11, 13, and 14, apply to you through the Terms of Service and the Acceptable Use Policy it incorporates.

Timelines and targets in this Policy are goals we aim to meet, not guarantees, except where a law that applies to us sets a deadline, in which case we follow that law. Nothing in this Policy increases wrxstack's liability. The disclaimers, the limitation of liability, and the other protections in the Terms of Service apply to everything this Policy describes. Nothing in this Policy limits a right you have under a law that cannot be limited by agreement.

23. Changes to this Policy

We may update this Policy to reflect changes to the AI features, to the model providers we use, or to the law. We will give notice where the law or the Terms require it, and the effective date at the top shows the current version.

24. Contact

For any question about this Policy or about AI in the Service, email contact@wrxstack.com. Related documents: Terms of Service, Acceptable Use Policy, Privacy Policy, Data Processing Addendum, GDPR Statement, California Notice, Sub-processor List, Data Deletion Policy, and Data handling.