Guardrly Mcp
Audit what your AI agent actually calls on Shopify, Meta Ads, and similar APIs so you catch risky or unexpected automation in production.
Overview
Guardrly MCP is a MCP server for the Operate phase that audits AI agent API calls to platforms like Shopify and Meta Ads without invasive app changes.
What is this MCP server?
- Non-invasive monitoring of AI agent API usage—no invasive hooks described in the manifest
- Audit coverage called out for Shopify, Meta Ads, and additional integrations
- PyPI package guardrly with stdio MCP transport (version 0.1.3)
- Requires GUARDRLY_API_KEY and HMAC_SECRET from app.guardrly.com settings MCP panel
- Optional GUARDRLY_API_URL override defaulting to https://api.guardrly.com
- Package and server version 0.1.3
- PyPI identifier guardrly with stdio transport
- Three environment variables documented: GUARDRLY_API_KEY (required secret), HMAC_SECRET (required secret), GUARDRLY_API_
What problem does it solve?
Your agent can change live store and ad accounts, but you lack a clear audit trail of every API call it makes on your behalf.
Who is it for?
Solo founders running Claude or Cursor agents on Shopify or Meta Ads who need operational visibility and accountability.
Skip if: Greenfield apps with no third-party agent integrations, or teams that only need static code review without runtime API auditing.
What do I get? / Deliverables
After registration, Guardrly-backed MCP tooling lets you monitor and review agent API activity against configured commerce and ads integrations.
- MCP-linked audit channel for agent API activity
- Configured connection to Guardrly API (default https://api.guardrly.com)
- Operational visibility for commerce and ads automations driven by agents
Recommended MCP Servers
Journey fit
Guardrly fits Operate because monitoring and proving what agents did matters after you ship agent workflows that touch live merchant and ad accounts. Monitoring subphase is the right shelf for non-invasive API call auditing that runs alongside production agent traffic.
How it compares
Runtime agent API audit MCP, not a web scraper or financial tool catalog.
Common Questions / FAQ
Who is Guardrly MCP for?
Builders and operators who automate Shopify, Meta Ads, or similar APIs with AI agents and need audited visibility into those calls.
When should I use Guardrly MCP?
Use it when agents are live against merchant or ads accounts and you want monitoring and audit evidence in Operate—or while hardening before launch in Ship.
How do I add Guardrly MCP to my agent?
Install the guardrly PyPI MCP package for stdio, set GUARDRLY_API_KEY and HMAC_SECRET from app.guardrly.com/settings, and optionally GUARDRLY_API_URL.