Can I use PolicyGuard in production today?
AI Agent Compliance APIs

Can I use PolicyGuard in production today?

3 min read

Getting started

Can I use PolicyGuard in production today?

No. PolicyGuard is a prototype built during a one-day hackathon on May 23, 2026. We do not have a production deployment yet, and we are not offering it as a paid service today. (Source: https://policyguard-site.vercel.app/#status, https://policyguard-site.vercel.app/#faq)

What is PolicyGuard?

PolicyGuard is a compliance API for AI agents acting on the open web. Given a proposed action and a target site, we return a verdict, a risk level, a short reason, a safe alternative when needed, and cited evidence from the actual policy text. (Source: https://policyguard-site.vercel.app/#what)

Who is PolicyGuard for?

PolicyGuard is for other AI agents, not humans clicking through a UI. We built it for developers who want autonomous software to avoid violating site terms of service. (Source: https://policyguard-site.vercel.app/#faq)

How it works

How do we use PolicyGuard?

You call our POST /evaluate endpoint with the proposed action and target site. We fetch the relevant policy page(s), ground the text, and return a structured decision in one HTTP round-trip. (Source: https://policyguard-site.vercel.app/#products)

What does PolicyGuard return?

We return one of three verdicts: allowed, blocked, or modify. The response also includes a risk level, a human-readable reason, a safe alternative if the action is blocked, and cited evidence with the source URL and quoted policy passage. (Source: https://policyguard-site.vercel.app/#what)

Why do we use x402?

x402 lets an agent pay per request without prior signup or a human provisioning credentials. That fits autonomous agents that may discover PolicyGuard at runtime and need to complete payment on their own. (Source: https://policyguard-site.vercel.app/#faq, https://policyguard-site.vercel.app/#products)

Limits and edge cases

What is cited.md?

cited.md is our public, growing corpus of grounded decisions. We publish every verdict there so future agents can search and cite prior answers instead of re-paying for the same lookup. (Source: https://policyguard-site.vercel.app/#products)

What happens if a site does not have a clear policy?

We do not guess. When the policy is ambiguous or missing, the verdict is modify or blocked, and the reason explicitly says that no clear authorization was found. (Source: https://policyguard-site.vercel.app/#faq)

Does PolicyGuard guarantee legal compliance?

No. We surface and cite the relevant policy text, but we do not replace legal review. The calling agent and its operator remain responsible for the action they take. (Source: https://policyguard-site.vercel.app/#faq)

Why not just hardcode compliance rules?

Hardcoded rules go stale when policies change, miss sites you did not anticipate, and provide no citation. PolicyGuard uses live policy fetch plus citations so decisions stay current and auditable. (Source: https://policyguard-site.vercel.app/#problem)