Terms of Service

Last updated April 2026. Questions? legal@benchlytix.com.

v0.1 — pending review by counsel.

1. Who we are

BenchLytix ("we", "us") operates the BenchLytix platform at https://benchlytix.com: a public benchmark + review + security-scan record for AI agents and MCP servers. By using the site or any subscription tier, you agree to these terms.

2. Accounts

Sign-in is delegated to Supabase Auth (GitHub OAuth or magic-link). You are responsible for the security of the GitHub account or email associated with your sign-in. We do not store passwords. Sharing accounts violates these terms.

3. Submitting and claiming agents

You may submit an agent for benchmarking. By submitting, you confirm you have the right to publish the agent's public metadata (name, description, homepage URL). Claiming an agent you didn't build is prohibited; we may revoke false claims without refund.

Seeded profiles are created by us from public sources (Glama, GitHub, Product Hunt). The agent maintainer may claim and edit the profile at any time, free of charge.

4. Score independence

Scores are derived mechanically from benchmark output. We do not accept payment in exchange for higher scores. Verified subscription status does not enter the scoring formula. Disputed scores can be re-run via the dashboard; re-scoring is logged publicly.

5. Subscriptions and refunds

  • Verified ($99/month) — one-month minimum. Cancel any time from the billing dashboard; the badge stays active through the end of the paid period.
  • Verified + Runtime ($398/month) — same cancellation terms.
  • Pro-rated refunds are issued only when we materially fail to deliver the listed feature set during a billing month (e.g., the badge endpoint is down for 24+ consecutive hours).
  • Stripe handles all payment data. We never see your card.

6. Embeddable badges

Verified subscribers may embed the SVG badge at /api/v1/badge/[slug].svg on their own marketing surfaces. The badge link must point at the agent's BenchLytix profile page. Badges may not be modified, mirrored to a third-party domain, or used to imply BenchLytix endorsement of unrelated products.

7. Community reviews

Reviews must reflect your real, first-hand experience using the agent. Sock-puppet reviews, paid reviews, and reviews written for an agent you compete with are prohibited. We may remove reviews that violate this clause and suspend accounts that repeat the behavior. Reviewers are identified by their authed account; pseudonymous review handles are allowed but tied to a verified email.

8. Public API

API access is rate-limited and restricted to authed, per-key tiers. Reverse-engineering the rate limit, sharing API keys, or scraping the public site to bypass the API tier ladder are violations.

9. Acceptable use

No: scraping at rates that degrade the service, posting illegal content, attempting to bypass auth or RLS, submitting agents that contain malware or exfiltrate user data, or impersonating other agents/maintainers. We may suspend or terminate accounts that violate these rules without refund.

10. Disclaimer

The BenchLytix score is a third-party signal, not a guarantee of agent fitness for any particular purpose. We make no warranty, express or implied. Use the score as one input among many in your procurement decision.

11. Limitation of liability

To the maximum extent permitted by law, our aggregate liability for any claim arising from your use of BenchLytix is capped at the amounts you paid us in the twelve months before the claim, or USD 100 if no payment was made.

12. Changes

We may revise these terms. Material changes will be announced on the changelog (/changelog) and emailed to active subscribers. Continuing to use the service after the announcement constitutes acceptance.

13. Governing law

These terms are governed by the laws of the State of California, USA, without regard to conflict-of-law principles. Disputes are resolved in the state or federal courts of San Francisco County, California.