Skip to content
G

Guardyx AI

Sign in / Sign up

Put an approval gate in front of your AI agent in 5 minutes.

Claude (desktop or claude.ai), Cursor, Meta Muse, Grok Bot or the agent you built: point it at Guardyx AI and every tool call is checked against your rules before it runs. Risky ones wait for you. Every decision is on a hash-chained record.

Every AI action passes through Guardyx.

Start freeRead the docs

Free: 5,000 tool calls a month through Guardyx, every governance feature, connectors included. No card, no expiry.

Approval requested

Held for a human before it runs.

High riskPending
Agent
Claude Desktop
Tool
GMAIL_SEND_EMAIL

Agent's stated reasonUnverified

You asked me to tell customers about the new Q3 pricing, so I’m emailing the full customer list.

Arguments

{
  "to": "ALL-CUSTOMERS@list.acme.com",
  "body": "Effective immediately, new Q3 pricing…"
}

Policy

Bulk external email to a distribution list: approval required.

Your AI agents aren't just chatting anymore.

Claude, Cursor, and the agents you build now act with your credentials, faster than you can read what they'll do.

If your agents only draft text and never call a real tool, you probably don't need this yet.

Still weighing it up?

Why GuardyxSee pricing

It only takes one.

  • Force-push over your main branch.
  • Email everyone in your contacts.
  • Delete files or a table you still needed.
  • Paste a secret key somewhere public.
  • Spend money with your card or API credits.
  • Cancel a meeting with your biggest customer.

Each happens in seconds, under your name, and you only find out afterward.

One gate for every agent you run.

Point every agent and MCP client at Guardyx. One control point, one record.

Your agents

  • Claude (desktop or claude.ai)
  • Cursor
  • Meta Muse
  • Grok Bot
  • ChatGPT (over MCP)

Guardyx

Policy, approvals, audit

hosted, or in your VPC

The systems they act on

  • Sales & email tools
  • Payments & ledgers
  • Code & repos
  • Production infrastructure
  • Internal data & records
Building your own? Adapters for CrewAI, AutoGen, Google ADK, OpenAI Agents SDK, Anthropic Claude Agent SDK, LangGraph & LangChain, OpenAI tool calling.

Connect your MCP clients.

Every client above signs in over OAuth 2.1, except ChatGPT, which connects directly. No adapter needed.

Connectors included.

Guardyx hosts the connectors and governs them. One bill, no separate connector subscription, no agent code to change.

Bring the agents you already run.

Native adapters for CrewAI, AutoGen, Google ADK, the OpenAI and Anthropic agent SDKs, and LangGraph, with LangChain tools governed through that same LangGraph adapter.

See how agents connect

Control at the moment of action.

Not a log you read afterward. The decision happens before the action runs.

Your agent proposes

GMAIL_SEND_EMAIL

to every customer on a list

Guardyx

Checks it against your policy

before anything runs. If the check fails, nothing runs.

  • Allowed

    In policy, so it runs now.

  • Held

    Waits for a human to approve.

  • Denied

    Never reaches the system.

Audit trail

Every decision recorded

hash-chained, including the ones that never ran.

Fails closed.

Guardyx fails closed by default: if the policy check cannot complete, the action does not run.

Keeps the receipts.

A hash-chained, tamper-evident record of what was allowed, held or denied.

Runs where you need it.

Hosted, or self-hosted in your own VPC or on-prem.

No code in your agent.

The gate lives in Guardyx, not in your agent loop. Approvals work for any MCP client the day you connect it.

This is the screen you'll use.

The real approvals queue: risky tool calls held before they run.

The Guardyx AI approvals queue: pending agent tool calls such as GMAIL_SEND_EMAIL and NOTION_CREATE_PAGE, each tagged high or critical risk and held for a human to review before they run.
An approval's detail panel in Guardyx AI, showing the held request and the agent's stated reason for it, marked unverified.

Open a held action, see why it stopped, approve or deny.

Batch review in Guardyx AI: several low-risk email actions selected together, with controls to approve or deny all of them at once.

Clear routine actions in one batch.

Prevent. Review. Prove.

Nobody wants to be the person explaining why an agent did something nobody approved. Guardyx does all three jobs, on every action.

  • Prevent

    No risky action runs without a policy decision.

    Intercept the action and check it against your policy.

  • Review

    You see only what matters.

    Hold the exceptions for your sign-off; let routine work through.

  • Prove

    Export the evidence when you need it.

    Run it, and log the decision as tamper-evident proof.

The Guardyx AI Runs Explorer with agent work grouped by MCP session: each session's calls, apps and time span, and a timeline of the tools the agent searched for and called, each with its risk and policy decision.

What your agent actually did last week: every run, every decision, exportable as a hash-chained evidence pack.

See the full platform

How it compares

What each tool does today, from its own site and docs.

  • Approvals enforced in the gateway, no agent code

    Guardyx
    Yes
    Composio
    Partial
    Mintive (formerly MintMCP)
    Yes
    Runlayer
    Partial
    Docker MCP Gateway
    No

  • Hash-chained audit log

    Guardyx
    Yes
    Composio
    No
    Mintive (formerly MintMCP)
    Partial
    Runlayer
    No
    Docker MCP Gateway
    No

  • Connectors + governance on one bill

    Guardyx
    Yes
    Composio
    Yes
    Mintive (formerly MintMCP)
    Yes
    Runlayer
    Not stated
    Docker MCP Gateway
    No

  • Works with Claude Desktop / Cursor over MCP OAuth 2.1

    Guardyx
    Yes
    Composio
    Partial
    Mintive (formerly MintMCP)
    Yes
    Runlayer
    Yes
    Docker MCP Gateway
    No

  • Unlimited agents on the free plan

    Guardyx
    Yes
    Composio
    Not stated
    Mintive (formerly MintMCP)
    No
    Runlayer
    Not stated
    Docker MCP Gateway
    Not stated

  • Start self-serve, no sales call

    Guardyx
    Yes
    Composio
    Yes
    Mintive (formerly MintMCP)
    Partial
    Runlayer
    No
    Docker MCP Gateway
    Yes

  • Free tier

    Guardyx
    Yes
    Composio
    Yes
    Mintive (formerly MintMCP)
    No
    Runlayer
    Not stated
    Docker MCP Gateway
    Yes
Sources, checked 28 Oct 2026

Using it at work?

Invite teammates and your workspace grows in place: same tenant, same audit chain, no migration. Paid plans add unlimited teammates, 90 days of searchable history, metering you can bill from, and SSO/SAML/SCIM. Enterprise adds self-hosting in your VPC, bring-your-own connectors and MCP servers, and regulator-ready exports.

See plansRequest Enterprise access

Start free in your browser.

Every AI action passes through Guardyx.

5,000 tool calls a month through Guardyx, every governance feature, connectors included, up to 3 teammates. No card, no expiry.

Start freeSee pricingRead the docs