Terms of Service
AgentProof Terms of Service
These terms explain what AgentProof is, how you may use it, and what it does and does not promise. They are written in plain English and describe how the product actually works today.
1. What AgentProof is
AgentProof is software that gives you access to run AI-agent readiness assessments and to build a versioned, in-workspace, evidence-backed compliance-readiness record. It is available two ways: the Compliance Readiness Pilot is a one-time paid engagement, and the Team plan is an ongoing monthly subscription. It provides compliance-readiness and evidence documentation support, with the readiness score as one secondary signal, based on the answers and evidence you provide.
2. What AgentProof is not
AgentProof is a decision-support aid. To be clear about what it is not:
- Not a certification or accreditation.
- Not regulatory or legal approval.
- Not legal advice or compliance advice.
- Not financial advice.
- Not insurance.
- Not an audit certification.
- Not a runtime monitor — it does not watch live agent traffic.
- Not a vendor endorsement.
Live behaviour testing is required before production use. A compliance-readiness record reflects design readiness from the information you supplied; it is not a guarantee of how an agent will behave in production.
3. Acceptable use
- Use AgentProof to assess agents you own or are responsible for.
- Provide accurate answers and evidence — the assessment is only as good as the information you give it.
- Do not misrepresent a compliance-readiness record as a certification, an approval, or a guarantee of an agent's behaviour.
- Do not attempt to disrupt the service, access another customer's data, or use AgentProof to break the law.
4. The assessment and records
- You supply answers and supporting evidence; AgentProof applies the question packs that match what your agent does.
- The output is a readiness score and band plus a written compliance-readiness record of findings, evidence gaps, blockers, and recommended next actions.
- Older records are kept immutable — AgentProof never edits or silently rewrites them. They can still be permanently deleted by you: self-service data deletion removes your records, and a workspace owner can delete the whole workspace.
- AgentProof scores design readiness. It does not watch your agent's live traffic and is not a runtime monitor.
5. Accounts and access
Pilot and team access are arranged directly with AgentProof: the one-time EUR 1,500 Compliance Readiness Pilot (up to 5 AI agents), the EUR 199 per month Team subscription (one workspace, up to 25 AI agents), and a custom Enterprise / private deployment option. The Pilot is a single fixed-fee engagement, not a subscription; the Team plan renews monthly until you cancel. You are responsible for keeping your sign-in credentials secure and for activity that happens under your account.
6. Availability and changes
We aim to keep AgentProof available on a reasonable-efforts basis, but we do not promise uninterrupted access. The assessment methodology can evolve over time; changes pass through a human review gate before they affect records, and historical compliance-readiness records stay immutable. You can review the public change history on our change log.
7. Responsibility and liability
AgentProof is a decision-support aid. You remain responsible for your agent's behaviour, for the decisions you make based on a compliance-readiness record, and for carrying out live behaviour testing before putting an agent into production. To the extent permitted by law, AgentProof is provided as-is and we are not liable for outcomes that depend on how you build, configure, or operate your agents.
8. Contact
Questions about these terms? Visit our help centre to get in touch. For legal correspondence about these terms, you can write to [email protected].
This is the current Terms of Service. It is not legal advice.
See also our Privacy Policy, Refund Policy, and Pricing.
Plain-English terms, written to match how AgentProof actually works.