Acceptable Use Policy

Last updated: October 10, 2026
Effective date: [Insert publication date]

This Acceptable Use Policy (“Policy”) applies to your use of Garagent, including agents, tasks, projects, content, skills, memories, and connected tools or services. Garagent is operated by ZAD Creation LLC (“we,” “us,” or “our”). This Policy forms part of our Terms of Use. Capitalized terms not defined here have the meanings given in those Terms. If there is a conflict, the Terms control.

You are responsible for your use of the Service and the instructions, permissions, integrations, and schedules you configure. These restrictions apply whether you act directly or use an agent, tool, or other automated process. You must not knowingly enable, facilitate, or attempt a prohibited use. Nothing in this Policy removes our responsibilities for our own conduct or limits rights that applicable law protects.

1. Prohibited uses

You must not use Garagent to:

Engage in unlawful activity or materially assist others in doing so, including unlawful transactions or evasion of applicable sanctions or export restrictions.
Infringe copyright, trademarks, trade secrets, privacy, publicity, or other rights.
Commit fraud, conduct phishing, distribute scams, steal identities, or deceptively impersonate a person or organization.
Send unsolicited bulk messages; disregard recipients’ opt-outs; fabricate reviews, testimonials, or engagement; or conduct deceptive campaigns.
Harass, threaten, extort, exploit, or abuse people; promote violence or terrorism; or facilitate human trafficking or sexual exploitation.
Create, obtain, solicit, or distribute child sexual abuse material or other content sexually exploiting minors, including synthetic or AI-generated depictions.
Create or distribute non-consensual intimate imagery, including sexual deepfakes, or use synthetic media for fraud, intimidation, or other prohibited conduct.
Create or distribute malware, steal credentials, or gain unauthorized access to accounts, devices, or networks.
Attack, disrupt, probe, or overload a system without authorization, or act outside the scope of authorized security testing.
Collect, disclose, infer, or use personal information without an appropriate lawful basis and any required permissions or notices.
Conduct unlawful or covert surveillance, stalk individuals, or expose private information to intimidate or harm them.
Develop or operate weapons or use the beta to control safety-critical systems where a failure could reasonably cause death, physical injury, or serious environmental harm.
Violate applicable terms or usage policies of an AI provider, connected platform, or other third-party service used through Garagent. You must comply with restrictions applicable to the provider and feature you select.

These restrictions concern the conduct and use involved. Lawful reporting, education, research, and analysis are permitted where they do not themselves involve prohibited conduct and comply with applicable provider restrictions.

2. Agents and automated actions

Grant agents only the permissions needed for their tasks. You must have authority to access each account, mailbox, file, system, or data source you connect, and to authorize the resulting processing and actions.

You must not instruct agents to bypass approval requirements, access restrictions, safety controls, or opt-out preferences. Do not use indirect instructions, imported skills, agent-to-agent delegation, or connected tools to evade these restrictions.

Verify intended recipients, destinations, and the scope of consequential actions. Use appropriate human oversight, monitor scheduled work, and stop or disable tasks when you become aware of harmful or unauthorized behavior. Where necessary, revoke access directly with the connected provider. Garagent does not guarantee that a submitted action can be canceled or reversed.

You must not use Garagent as the sole basis for decisions with legal or similarly significant effects on people, including decisions about employment, credit, healthcare, housing, insurance, or essential services. Appropriate independent human review, lawful decision criteria, and any legally required safeguards must be in place. Human review does not make an otherwise unlawful use permissible.

Do not misleadingly represent AI-generated material as independently verified, professionally certified, or created or endorsed by a particular person. Provide AI disclosures where required by law or the platform on which you publish.

For communications and marketing, you are responsible for applicable consent requirements, accurate sender identification, required disclosures, and honoring opt-outs. Permission to connect a mailbox or messaging account does not by itself authorize contacting every recipient.

3. Protecting information

Only upload or connect information you are authorized to use and disclose. Relevant content may be sent to AI providers, connected tools, or intended recipients to complete tasks, as described in our Privacy Policy.

Do not place passwords, private keys, access tokens, or other secrets in ordinary chats or files. Use designated credential settings where available. Do not use exposed credentials belonging to another person.

During the beta, do not process complete payment-card details, government identification numbers, protected health information, or similarly sensitive regulated information unless we have expressly agreed in writing to support that use and appropriate safeguards are in place.

Do not submit an employer’s, customer’s, or other party’s confidential information if doing so would violate a duty of confidentiality or other restriction. Minimize unnecessary personal information and review content before sending or publishing it.

4. Misuse of the Service

You must not:

Bypass authentication, account restrictions, credit limits, rate limits, billing controls, or other security and usage controls.
Access another user’s account, content, workspace, or administrative features without authorization.
Exploit vulnerabilities or continue accessing information accidentally exposed to you.
Interfere with the Service through excessive requests, uncontrolled agent loops, malicious files, or other disruptive activity. If an unintended loop occurs, take reasonable steps to stop it and report it.
Create multiple accounts, use false identities, or rotate accounts or credentials to evade restrictions or obtain unauthorized credits.
Share or transfer invite-only access to unapproved users.
Resell, sublicense, or make access available to others without our written agreement.
Copy or commercially exploit proprietary software or materials except as authorized, or reverse engineer the Service except to the extent permitted by applicable law.

These restrictions do not prevent permitted use of your own Input or Output under the Terms.

5. Security research

If you discover a possible vulnerability or unintended exposure, stop activity that could expose additional information or disrupt the Service and report it promptly. Do not access additional data to demonstrate the issue, modify or delete information, retain unnecessary copies, or disclose private user information.

Obtain our written authorization before active security testing against the Service. Authorization must identify the systems, methods, and scope permitted. Our authorization does not extend to third-party systems, AI providers, or connected services without their permission.

This Policy does not itself establish a bug bounty or authorize testing. It does not restrict disclosures protected or required by applicable law, including lawful reports to regulators or appropriate authorities.

6. Enforcement

We may investigate suspected violations and take proportionate action based on their severity, risk, and recurrence. Actions may include warnings, tool restrictions, pausing agents or tasks, removing or restricting content, and suspending or terminating access under the Terms.

We may act immediately where reasonably necessary to prevent harm, protect users, secure the Service, or comply with law. We will provide notice and an explanation where reasonably practicable, unless doing so would compromise an investigation, security, or legal obligations. Where appropriate, we may allow an issue to be corrected before restoring access.

We may process relevant account records and content to investigate or enforce this Policy, subject to our Privacy Policy and applicable law. We may preserve or disclose information where legally required or otherwise lawfully permitted for the purposes described in that policy.

We do not promise to inspect every task or detect every violation. Failure to act on a particular incident does not authorize similar conduct or waive later enforcement.

If you believe an action was mistaken, contact hello@getgaragent.com with relevant details. We will review the request and communicate the outcome as appropriate, including any review or appeal rights required by law.

7. Reporting

Report misuse, unintended agent behavior, or security concerns to hello@getgaragent.com.

Include a description, relevant dates, and task or account identifiers where available. Do not send passwords, access tokens, unnecessary personal information, or copies of illegal content. For suspected child sexual abuse material, describe its location without downloading or forwarding it.

Garagent’s support mailbox is not an emergency service. Contact appropriate emergency services if someone faces immediate danger.

8. Changes

We may update this Policy as the Service develops or new risks arise. We will revise the date and provide notice of material changes as described in our Terms. Changes apply prospectively, and we will obtain renewed acceptance where required by law. Nothing in this section prevents immediate enforcement of existing rules or action required by law.

9. Contact

Garagent is operated by ZAD Creation LLC.

Questions, reports, and requests for review: hello@getgaragent.com.