Legal

Terms of Service

Last updated: March 20, 2026

1. Acceptance

By accessing Rhumb (“the Service”), operated by Supertrained Inc., you agree to these Terms of Service. If you disagree, please do not use the Service.

2. Nature of scores

AN Scores are opinions based on documentation analysis and, where available, runtime testing . They are not warranties, guarantees, or certifications of any kind. Scores reflect our assessment of how well a service's published documentation and observed behavior suggest it will work for autonomous AI agents.

Scores may change as documentation is updated, runtime evidence is collected, methodologies are refined, or disputes are resolved. We make reasonable efforts to be accurate but do not guarantee the correctness of any individual score.

3. Service tiers and access

Rhumb offers multiple tiers of access:

  • Browse (free, no auth): Score data, service profiles, failure modes, leaderboards, and search are available without authentication.
  • Execution rails (paid as used): Authenticated execution may use governed API key or wallet-prefund on X-Rhumb-Key, while provider-controlled paths use BYOK or Agent Vault where supported. Discovery remains free, and current pricing and markup terms are published at /pricing.
  • x402 zero-signup (paid): Agents with x402 capability may use execution endpoints without an account by including a valid USDC payment proof with each request.

We reserve the right to modify access rails, pricing, and billing terms with reasonable notice.

4. Managed capabilities and proxy services

Rhumb offers managed capability execution, where we proxy API calls to third-party services on your behalf using credentials we manage. When using managed capabilities:

  • • You are responsible for compliance with the upstream service's terms of service
  • • Rhumb acts as a technical intermediary, not as the provider of the upstream service
  • • For Rhumb-managed execution rails, we apply the current markup disclosed at /pricing; BYOK or Agent Vault provider-controlled routes do not add markup to the credential itself
  • • We are not liable for upstream service outages, errors, or changes to their APIs
  • • We reserve the right to suspend managed capability access for any account engaged in abusive usage

5. Acceptable use

You agree not to use the Service to:

  • • Send unsolicited messages (spam) via managed capabilities
  • • Engage in phishing, fraud, or other deceptive practices
  • • Attempt to access other users' data, keys, or sessions
  • • Circumvent payment, rate limiting, or authentication mechanisms
  • • Use managed credentials for purposes unrelated to your legitimate application
  • • Systematically overload the API to degrade service for others

Violation of these terms may result in immediate suspension of access without refund.

6. Payments and refunds

Stripe prepaid balance: Funds added via Stripe are prepaid credits. Credits are non-refundable once consumed by capability executions. Unused credits may be refunded within 30 days of purchase by contacting billing@supertrained.ai.

x402/USDC payments: USDC payments are on-chain and final. Refunds for x402 payments are processed at our discretion and returned to the originating wallet address.

7. API and service disclaimer

The API and all services are provided as-is, with:

  • • No guaranteed uptime or service level agreement (SLA)
  • • No guaranteed response time or rate limit protections
  • • No commitment to backward compatibility of response schemas (we aim for stability but reserve the right to evolve the API)

8. Data licensing

  • Score data (AN Scores, failure modes, leaderboards): Licensed under Creative Commons Attribution 4.0 International (CC BY 4.0) . You may use, share, and adapt the data for any purpose, including commercial use, provided you give attribution.
  • Source code: Licensed under the MIT License .
  • Service guides and blog content: © 2026 Supertrained Inc. All rights reserved unless otherwise noted.

9. Dispute process

If you believe a score is inaccurate, you may file a dispute via:

We commit to reviewing every dispute and responding within 5 business days. Outcomes of public disputes are published on GitHub. We reserve final editorial discretion on all scores.

10. Limitation of liability

The Service is provided “as is” without warranty of any kind, express or implied. Supertrained Inc. shall not be liable for any damages arising from the use of AN Scores, failure mode data, managed capability executions, proxy services, or any other information or functionality provided through the Service.

This includes but is not limited to: decisions made based on AN Scores, tool selection influenced by leaderboard rankings, business impact from published failure mode data, costs incurred through managed capability execution, or upstream service failures during proxied calls.

11. Changes to terms

We may update these terms. Material changes will be announced in our changelog and noted by updating the “Last updated” date. Continued use of the Service after changes constitutes acceptance.

12. Governing law

These Terms are governed by and construed in accordance with the laws of the State of Florida, United States.

13. Contact

For questions about these terms:
legal@supertrained.ai

Supertrained Inc.
7901 4th St N STE 300
St Petersburg, FL 33702