NerdPress® Guided Agent Connector Terms

1. What this service does.

Our NerdPress Guided Agent Connector (the “Service”) lets you connect an AI agent you already use (such as OpenAI’s ChatGPT or Anthropic’s Claude) to your WordPress site. When you connect an agent through the Service, you authorize that agent to access and act on your WordPress site on your behalf. Depending on the permissions you grant, a connected agent may read your site’s content, media, comments, users, and settings, and may make changes to them — including creating, editing, publishing, and deleting content; changing site, plugin, and theme settings; and installing, activating, deactivating, or updating plugins and themes.

The Service does not give a connected agent your WordPress password, your hosting credentials, or your billing information. Beyond that, however, you should assume a connected agent can reach anything your site makes available to it. Depending on the permissions you grant and the plugins you run, that may include customer records, orders, form submissions, and other sensitive data. What any individual plugin exposes is determined by that plugin, not by us, and may change as the plugin is updated. Please grant access with that in mind.

We built in guardrails and controls designed to make working with an AI agent safer than giving one unrestricted access to your site — for example, we may warn on or block certain actions, require confirmation before particularly sensitive ones, limit what a connection can reach, and apply guidance drawn from our experience keeping WordPress sites healthy. These protections will change over time, and we may add, adjust, or remove any of them as we learn what works. “Safer” does not mean risk-free or guaranteed: our protections are a best effort, not a promise about any specific outcome, and they may not catch every risky action or may behave unexpectedly. In addition, the Service is software — our own code and infrastructure may contain errors, may fail or be unavailable, and could cause problems on your site, including lost or damaged content, independent of anything an AI agent does. Your remedies and our responsibility for all of the foregoing are governed by Sections 3 and 8.

If you maintain a staging (test) copy of your site, you may connect it through the Service and try out changes there before applying them to your live site. We recommend testing significant or sensitive changes on a staging copy first where practical. Staging is an option you control; using it is your choice, and — like our other protections — it reduces risk but does not eliminate it or guarantee any particular result.

2. You’re in control.

We provide tools to manage your connections — including setting and changing what each connection is allowed to do, revoking a connection, and reviewing what has been done through one. The specifics of these tools will change as the Service develops. Revoking a connection stops further activity through it, but it does not reverse actions that have already been taken.

A “connection” includes any AI agent, application, or third-party platform authorized to act through the Service — whether a person or an automated system is issuing the instructions. A connection may be authorized by anyone with sufficient access to your WordPress site, which can include your team members and contractors as well as you. Managing who has that level of access is your responsibility, and any connection authorized by those people is treated as authorized by you. You are responsible for every connection authorized on your site and for everything done through it.

If you authorize a connection on behalf of a business or another person, you are agreeing to these terms for that business or person as well as yourself, and you represent that you have authority to do so.

We do not access or store your passwords or payment card details.

Suspension. We may suspend, limit, or disable any connection, or the Service as a whole, at any time and without prior notice, if we believe it is necessary to protect the security, integrity, or proper operation of the Service, our systems, or our customers. We are not obligated to monitor for, detect, or prevent any particular problem, and our ability to suspend the Service is a protection we may use at our discretion — not a guarantee that we will detect or stop any given issue.

3. PROVIDED “AS IS”

WE PROVIDE THE SERVICE ON AN “AS IS” AND “AS AVAILABLE” BASIS. WE DO NOT PROMISE THAT OUR GUARDRAILS OR ACTION LOG WILL PREVENT, DETECT, OR REVERSE ANY PARTICULAR ACTION TAKEN BY AN AI MODEL, OR THAT THEY WILL OPERATE WITHOUT ERROR OR OMISSION OR ALWAYS BE AVAILABLE. WE DO NOT PROMISE THAT ANY OUTPUT GENERATED BY YOUR CHOSEN AI PROVIDER WILL BE ACCURATE, APPROPRIATE, SAFE, OR NON-INFRINGING. YOU ACKNOWLEDGE THAT AI MODEL BEHAVIOR IS PROBABILISTIC AND THAT NO SAFEGUARD CAN GUARANTEE A SPECIFIC OUTCOME. TO THE FULLEST EXTENT PERMITTED BY CALIFORNIA LAW, WE DISCLAIM ALL IMPLIED WARRANTIES, INCLUDING MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NON-INFRINGEMENT.

Any description of the Service’s features — anywhere, including in our marketing, website, or documentation — is for general information only and does not create any representation, warranty, or promise beyond what is expressly stated in these terms.

4. Backups and recovery.

Backups of your site are provided as part of your subscription plan, as described in the Support Plan Terms. If something goes wrong through the Service — whether caused by a connected agent or by our own tooling — we will help you recover, using reasonable efforts. Wherever practical, we will aim to restore only the affected content and leave the rest alone; in some cases, however, the only route back is a full restore to an earlier point in time, which will also reverse any other changes made since that point. Where we are able to offer a way to reverse a particular action, we will, but not every action can be reversed, and revoking a connection does not reverse actions already taken. Some information used to support recovery may be stored on your own site, and its availability depends on your site remaining intact; deleting or losing your site may also remove the means to reverse a change. Because anyone with sufficient administrative access to your site can authorize a connection, controlling who holds administrator access — and reviewing your connection list periodically — is an important part of keeping your site safe.

None of this is a guarantee. As stated in the Support Plan Terms, we cannot guarantee that a backup will always be available, current, or complete, and recovery is assistance we provide on a reasonable-efforts basis — not a promise that any given change can be reversed cleanly or that data can be fully recovered. As a general best practice for any website — independent of the backups included with your plan — we recommend keeping your own current backups as well, so you always have more than one way to restore. You remain responsible for reviewing changes made through the Service before you rely on or publish them, and for the security of your own AI provider accounts, API keys, and credentials, and all activity conducted through them. Our responsibility for backup and recovery is subject to Sections 3 and 8.

5. Your AI provider is your own.

You select, contract with, and pay your AI provider (such as OpenAI or Anthropic) directly. That provider is not our subcontractor, agent, or vendor. We do not control, monitor, endorse, or accept responsibility for any AI provider’s model behavior, availability, pricing, data-handling practices, or terms, and you are responsible for complying with your AI provider’s terms and for any liability arising from that provider’s acts or omissions. The same applies to any third-party plugins, themes, applications, or platforms you run or that you direct a connection to install, activate, or update: we do not control them, and what any of them exposes, changes, or does to your site — and what data it makes available to a connected agent — is determined by that third party, not by us, as described in Section 1. If your connected AI communicates with your site’s visitors in California, you are responsible for giving any “you’re talking to a bot” disclosure that California law requires (Cal. Bus. & Prof. Code §§ 17940–17943).

6. Privacy, logging, and consent.

We do not retain or read the prompts you send to your connected AI agent. To provide, secure, operate, and improve the Service — and for audit logging, abuse prevention, and support — we do keep logs of the requests and actions taken through your connection (for example, a record that a given change was made), along with session metadata, IP addresses, and user identifiers. We keep these connection logs for a limited period — currently about 30 days, although some technical or server-side logs may be kept longer — and these periods may change over time. We will not use this information to train, fine-tune, or develop any AI model, or to build any product or service other than the Service we provide to you. You represent and warrant that, before routing any individual’s data through the Service, you have provided all required notices and obtained all required consents and lawful bases from every individual whose personal information may be transmitted through, or retained in backups or logs by, the Service — including your own personnel, contractors, and your site’s visitors/end users. You agree not to transmit through the Service any protected health information, other special-category/sensitive data, or payment card data for which you have not established a lawful basis and appropriate safeguards. We act as your service provider with respect to this data, and our handling of it is governed by the Data Processing Addendum (Exhibit A), which is incorporated into these terms. If you or your site’s visitors are located in the EEA or United Kingdom, supplemental data-protection terms (including a cross-border transfer mechanism) will apply and are forthcoming, as described in Exhibit A; in the meantime you are responsible for determining whether your use of the Service is lawful for those individuals and for limiting the personal information you route accordingly. We are not a HIPAA business associate and will not knowingly accept PHI.

7. Acceptable use.

You agree not to use the Service: in connection with any cryptocurrency, crypto-asset, token issuance, digital-asset custody, or exchange activity; to transmit unlawful, infringing, or defamatory content; to circumvent the Service’s guardrails or audit logging; in violation of applicable sanctions or export-control laws; to install, deploy, or distribute any plugin, theme, code, or content for which you do not hold the necessary rights or licenses; or for any purpose prohibited by applicable law. We may suspend your access to the Service immediately upon reasonable suspicion of a breach of this section, consistent with our Acceptable Use Policy.

8. LIMITATION OF LIABILITY.

EXCEPT FOR LIABILITY THAT CANNOT BE LIMITED UNDER APPLICABLE LAW, OUR TOTAL AGGREGATE LIABILITY ARISING OUT OF OR RELATING TO THE SERVICE WILL NOT EXCEED THE FEES YOU PAID US UNDER YOUR SUBSCRIPTION IN THE TWELVE (12) MONTHS PRECEDING THE EVENT GIVING RISE TO THE CLAIM. IN NO EVENT WILL WE BE LIABLE FOR LOST PROFITS OR REVENUE, LOSS OF GOODWILL, LOST OR CORRUPTED DATA, BUSINESS INTERRUPTION, OR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, EXEMPLARY, OR PUNITIVE DAMAGES, WHETHER OR NOT FORESEEABLE AND REGARDLESS OF THE THEORY OF LIABILITY.

The Service is not subject to any service-level agreement, uptime commitment, service credit, or liquidated-damages provision; any service-level commitments in the Support Plan Terms apply solely to our website support and maintenance services and do not extend to the Service. Nothing in these terms limits or excludes any liability that cannot be limited or excluded under California law — including liability for fraud, willful misconduct, gross negligence, or personal injury. The cap in this section does not apply to your indemnification obligations in Section 9, your breach of Section 6 or 7, or amounts you owe us for your subscription.

9. Indemnification.

We will indemnify you against third-party claims to the extent arising solely from our own negligent act, error, or omission in providing the Service, subject in all cases to the limitation of liability in Section 8. You will indemnify us against all claims, damages, and reasonable expenses (including reasonable legal fees) arising from: (a) the instructions, prompts, or content you or your authorized connections submit through the Service; (b) your breach of the privacy and consent commitments in Section 6; (c) your failure to maintain backups or safeguard your credentials; (d) your use of the Service in violation of Section 7, applicable law, or your AI provider’s terms; and (e) any infringement claim arising from content generated or actions taken through your connections. Our indemnity obligations, and the conduct of any claim, are subject to our insurers’ right to control the investigation, defense, and settlement of any claim for which we may be liable; neither party will settle a claim in a way that imposes liability or an admission on the other without that party’s prior written consent (not to be unreasonably withheld).

10. Notice and cooperation.

You will notify us in writing within forty-eight (48) hours of becoming aware of any incident, claim, demand, or regulatory inquiry relating to the Service, and will cooperate reasonably with us and our insurers in investigating and defending it. You will not admit liability on our behalf, settle any claim implicating us, or make any public statement attributing fault to us, without our prior written consent. Late notice will not by itself extinguish a claim, but to the extent your delay actually prejudices us or our insurers, our obligations may be reduced accordingly.

11. Independent contractors.

The parties are independent contractors. Nothing in these terms creates a partnership, joint venture, agency, fiduciary, or employment relationship, and neither party has authority to bind the other.

12. Governing law and dispute resolution.

These terms are governed by the laws of the State of California, without regard to its conflict-of-laws principles. First, the parties will attempt in good faith to resolve any dispute informally and, if that fails, through mediation. If mediation does not resolve it, any dispute will be finally resolved by binding arbitration seated in Los Angeles County, California, administered by JAMS under its rules then in effect; the arbitrator decides questions about the scope and enforceability of this arbitration agreement. Class-action waiver: disputes are resolved individually, and not as a class, collective, or representative action. Carve-outs: nothing here prevents either party from (a) bringing a qualifying claim in small-claims court, (b) seeking injunctive relief in court to protect confidential information or intellectual property, or (c) seeking public injunctive relief, which remains available in court.

Opt-out: you may opt out of arbitration within 30 days of first accepting these terms by emailing support@nerdpress.net; opting out will not affect the rest of these terms. If you are located outside the United States, mandatory local consumer-protection and forum rules may apply to you regardless of the governing-law and arbitration provisions above, and nothing in these terms waives rights that cannot be waived under the law that applies to you.

13. Eligibility and availability.

The Service is offered to our subscription customers. Which features are available may vary by subscription tier, and we may add, change, gate, suspend, or remove features over time. We may also make the Service available through authorized partners under separate terms; if you access the Service through a partner, additional or different terms from that partner may apply to your relationship with them.

14. How these terms fit together, and what survives.

These terms are part of, and incorporate, the Support Plan Terms and Exhibit A. If they conflict with the Support Plan Terms, these terms control for the Service only; the Support Plan Terms’ boilerplate (including notices, assignment, force majeure, severability, and entire-agreement provisions) otherwise applies. Sections 3, 4, 5, 6, 7, 8, 9, 10, 11, and 12 survive termination. If any provision is held unenforceable, it will be limited or severed to the minimum extent necessary so that the remaining provisions remain in full force.