Last updated: March 9, 2026
These terms govern your use of AgentCheck at agentcheck.care and agentcheck.clinic.
AgentCheck tests AI agents and chatbots for behavioral quality, security vulnerabilities, and brand alignment. We send synthetic test conversations to your bot and analyze the responses. We deliver a scored diagnostic report.
AgentCheck is a diagnostic tool, not a guarantee. A passing score does not certify that your bot is free of all vulnerabilities. A failing score identifies specific issues we detected — other issues may exist that our tests do not cover.
You must have authorization to test any bot you submit to AgentCheck. Do not submit bots you do not own or operate, or bots you do not have permission to test.
You are responsible for any API keys you provide. We recommend creating temporary API keys with spending limits for use with AgentCheck. We are not responsible for charges incurred on your API account during testing.
You must not use AgentCheck to:
Free scans are limited to 30 per week across all users. Individual users are limited to 2 free scans per week.
Paid checkups are charged via Stripe at the price shown at the time of purchase. Prices may change; existing purchases are not affected.
If your checkup fails due to our error (our service crashes, not your bot being unreachable), contact us for a full refund. We do not offer refunds for low scores — a low score means the product worked.
If your bot is unreachable during the checkup, we will attempt to connect for up to 5 minutes. If connection fails, you can retry with the same payment (the checkup ID remains valid).
You own your system prompts, bot configurations, and any data you submit. We do not claim ownership of your data.
The report format, scoring methodology, test modules, and diagnostic framework are our intellectual property. You may share your report with others (the magic link is shareable) but may not reproduce our methodology or create derivative products.
We own anonymized, aggregate statistics derived from scans. These contain no individually identifiable information.
5.1 AgentCheck is provided "as is" without warranty of any kind, express or implied.
5.2 We do not guarantee that our tests will find all vulnerabilities in your bot. New attack vectors emerge constantly.
5.3 We are not responsible for actions taken based on our reports. If our report recommends a fix and that fix causes issues with your bot, that is your responsibility to test before deploying.
5.4 We are not a compliance certification body. Our reports do not constitute legal, regulatory, or compliance certification.
To the maximum extent permitted by law, AgentCheck's total liability for any claim arising from use of the service is limited to the amount you paid for the specific checkup giving rise to the claim. We are not liable for indirect, incidental, or consequential damages.
We aim for high availability but do not guarantee uptime. We may perform maintenance, update tests, or modify scoring methodology at any time. We will try to provide advance notice for changes that materially affect reports.
We may refuse service to anyone at our discretion. You may stop using the service at any time. Existing reports remain accessible via their magic links.
These terms are governed by the laws of the State of Delaware, United States. Disputes will be resolved in the courts of Delaware.
We may update these terms. Continued use after changes constitutes acceptance.