These Terms of Service govern your access to and use of vr.dev (the "Service"), including the website, API, Python SDK, verifier registry, and documentation. By using the Service, you agree to these Terms and our Privacy Policy.

vr.dev provides verifiable reward functions for AI agent evaluation and training. The Service includes a hosted verification API, an open-source Python SDK, and a public registry of verifier and skill specifications.

• You must be at least 16 years old (or the age of digital consent in your country).
• You are responsible for your account, API keys, and all activity under them.
• Keep your API keys secure. Do not share them publicly or embed them in client-side code.
• Provide accurate information and keep it up to date. We may refuse or reclaim usernames.

API keys grant access to the hosted verification API. You are responsible for all usage associated with your keys. If you believe a key has been compromised, revoke it immediately via the dashboard.

• Do not share API keys with unauthorized parties.
• Do not use API keys to circumvent rate limits (e.g., rotating keys to exceed quotas).
• We may revoke keys that violate these Terms or exhibit abusive patterns.
• Free-tier usage is subject to rate limits. Exceeding limits may result in throttling or suspension.

You retain ownership of content you contribute to the Service, including verifier implementations, skill definitions, fixture sets, and documentation.

By contributing verifiers or skills to the public registry, you grant us a worldwide, non-exclusive, royalty-free license to host, store, reproduce, modify, publicly display, and distribute your Contributions as part of the Service. This includes making them available to other users via the API, SDK, and registry.

You represent and warrant you have all rights necessary and that your Contributions do not infringe others' rights.

The vr.dev Python SDK is distributed under its own open-source license. Your use of the SDK is governed by that license in addition to these Terms. The hosted API is a separate service subject to these Terms regardless of whether you use the SDK to access it.

When you use the verification API, we may store verification results, evidence artifacts, and usage metadata to operate the Service. This includes:

• Verification inputs (completions, ground truth) for the duration of the API request.
• Verification results and evidence artifacts, retained according to your plan's retention period.
• Usage metrics (request counts, latency) for billing and service improvement.

Evidence artifacts are stored with content-addressed hashing (SHA-256) for tamper-evidence. We do not use your verification inputs to train models or share them with third parties unless required by law.

• Do not violate laws or regulations.
• Do not use the Service to develop, train, or evaluate AI systems intended to cause harm.
• Do not submit malicious inputs designed to exploit verifiers, corrupt evidence, or manipulate reward signals.
• Do not attempt to reverse-engineer, extract, or replicate proprietary verifier implementations through API probing.
• Do not use the Service to generate misleading benchmark results or fabricate evaluation evidence.
• Do not scrape or harvest data without our consent (standard search indexing via robots.txt is allowed).
• Do not impersonate others or misrepresent affiliations.
• Do not circumvent security features, rate limits, or access controls.

The Service is designed for both human and AI agent access. Automated systems (including AI agents, CI pipelines, and training loops) may use the API subject to the same rate limits and acceptable use policies as human users.

• Automated systems must authenticate with valid API keys.
• You are responsible for the actions of any automated system using your API keys.
• We may apply additional rate limits to automated access patterns to protect service stability.

We use cookies and similar technologies to operate the Service, remember preferences, measure performance, and improve user experience. Some cookies are essential; others are for analytics. By using the Service, you consent to our cookie use as described here and in our Privacy Policy. Where required by law, we will present a cookie banner and obtain consent for non-essential cookies.

By using the Service, you agree to our processing of your information to provide and improve the Service.

The Service integrates with third-party tools and frameworks (including training libraries, agent frameworks, and MCP-compatible systems). We are not responsible for third-party content, policies, or practices. Use of third-party integrations is at your own risk.

Prices, features, and terms for paid tiers will be presented at purchase and may change with notice where required. Subscriptions renew unless canceled as instructed. Usage-based charges (per-verification overages) are billed at the rates displayed on your plan page. Taxes may apply.

The Service, including trademarks, logos, website code, and proprietary API infrastructure, is owned by us or our licensors and protected by law. Verifiers and skills in the public registry are licensed under their respective open-source licenses. Except for those licenses and the limited license to use the Service, no additional rights are granted.

If you believe content infringes your copyright, email us including:

• Identification of the copyrighted work and the allegedly infringing material (URL).
• Your contact info and a statement under penalty of perjury of your good-faith belief.
• A statement you are authorized to act, and an electronic signature.

You may stop using the Service and revoke your API keys at any time. We may suspend or terminate your account if you violate these Terms or if necessary to protect the Service or users. Upon termination, your API keys will be revoked and evidence artifacts will be deleted according to retention policies. Certain provisions survive termination.

The Service is provided "as is" and "as available" without warranties of any kind, express or implied. We do not warrant that verification results are accurate, complete, or suitable for any particular purpose. Verification results are tools for evaluation, not guarantees of agent behavior. You are responsible for how you use verification outputs in training, evaluation, and deployment.

To the maximum extent permitted by law, we and our affiliates are not liable for indirect, incidental, special, consequential, exemplary, or punitive damages, or loss of profits, data, or goodwill. This includes damages arising from reliance on verification results in AI training or deployment. Our total liability will not exceed the greater of (a) amounts you paid to us in the 12 months before the claim, or (b) $100.

You agree to indemnify and hold us harmless from any claims, damages, losses, and expenses (including reasonable attorneys' fees) arising from your use of the Service, your Contributions, AI systems trained or evaluated using the Service, or your violation of these Terms or law.

We may modify the Service and these Terms. We will post updates with a revised date. Continued use after changes means you accept the updated Terms.

Questions? Email us.