Guardrly Mcp

Guardrly is a local monitoring layer that records and alerts on the API calls an AI agent makes to platforms such as Shopify and Meta Ads.

What you can do

Intercept and record API calls an agent makes to external platforms such as Shopify and Meta Ads, producing a human-readable audit trail of what was done, when and why (no tools are listed in the manifest). Alert on dangerous operations before damage is done. Generate structured operation logs that can serve as appeal evidence for banned accounts. Scrub PII locally before any data reaches the cloud.

When to use it

When you let an AI agent act on external commerce or advertising platforms and want visibility into its calls, early warnings on risky operations, or an audit record you could use if an account is banned.

Requirements

Runs as a local process on the user's machine (desktop only) from the PyPI package guardrly over stdio. Requires a Guardrly account and two secrets: GUARDRLY_API_KEY and HMAC_SECRET, both from the Guardrly settings page. GUARDRLY_API_URL optionally sets the API endpoint. Network access to the Guardrly service.