Enterprise agent control plane

Govern every AI agent action before it touches the real world.

Guardyx AI is the runtime control plane for AI agents. Route every tool call through policy, approvals, audit, and observability — without rewriting your existing agents.

Request early access See a governed tool call

What happens after signup

1. Create your workspace.
2. Complete guided setup for agent, tool, and policy.
3. Run your first governed tool call.

Governed execution

How Guardyx AI mediates agent actions.

Agent actionsalesforce.update_opportunity

Policy verdictapproval required

Reviewersales manager

Trace statuscaptured

Why review

External write to a production CRM record, above the auto-approve threshold.

Payload

{
  "opportunity_id": "0064x00000Lp2yQ",
  "stage": "Closed Won",
  "amount": 250000
}

Connect your agents two ways — no rewrite either way.

Primary

Point a client at Guardyx (MCP)

Already using Claude Desktop, Cursor, or a custom MCP client? Point it at a Guardyx endpoint and every tool call it makes is governed — policy, approvals, and audit applied automatically. Setup is a config change, not a migration.

SDK

Wrap your own agent

Running a LangGraph or custom Python agent? Route its tool calls through the Guardyx SDK and get the same control plane around code you already have.

Explore the Platform

The runtime path inside Guardyx AI — how a tool call moves from interception to execution, with every decision captured.

Intercept

Route every tool call through Guardyx AI before it executes.

Evaluate

Run policies using context like user, tool, payload, and risk.

Approve

Require human approval for sensitive actions.

Trace

Capture inputs, decisions, and outcomes for every run.

Execute

Perform the action only after policies and approvals pass.

Replay

Coming soon

Re-run scenarios to debug behavior and validate guardrails.

Execution Flow

Without Guardyx AI

This sequence executes immediately — no policy checks, no approvals, no audit trail. You'd also need to build and maintain the control infrastructure yourself.

Agent decides to initiate a $250,000 vendor payment.

Call is routed through Guardyx AI instead of directly executing.

Policy check runs and flags the transaction as high-risk.

Execution is blocked and routed for finance approval.

Finance approver reviews details and approves the payment.

Payment executes and a full audit trail is recorded.

Execution Log

blocked

run_id: 9f3c2a1
agent: sales-agent
tool: salesforce.update_opportunity
status: blocked
reason: requires_approval
policy: external_write_requires_review

What You Get

Skip building it yourself

Keep your existing agents and route their actions through Guardyx — instead of building approval, audit, and policy infrastructure from scratch.

Real-world Use Cases

Sales Ops Automation

Agents update CRM records, create opportunities, and trigger follow-ups — with approvals for high-value changes.

Customer Support

Agents draft replies, issue refunds, or modify accounts — with safeguards before external communication.

Finance Workflows

Agents process invoices, initiate payments, and reconcile records — with strict policy enforcement and audit logs, without building custom approval and audit systems.

Internal Copilots

Employees use agents to query data and trigger actions across internal systems without direct-access risk.

Research Agents

Agents gather data, summarize insights, and optionally act on findings — all actions governed and traceable.

DevOps Automation

Agents deploy services, modify configs, or run scripts — gated by approval and full execution visibility, without building internal guardrail tooling.

Minimal Integration

Keep your agents. Route their actions through Guardyx instead of building a control plane from scratch.

MCP client (e.g. Claude Desktop / Cursor) — add the Guardyx endpoint to your client config:

{
  "mcpServers": {
    "guardyx": {
      "command": "npx",
      "args": [
        "-y", "mcp-remote",
        "https://your-guardyx-host/v1/mcp",
        "--header", "X-API-Key:gx_live_YOUR_KEY"
      ]
    }
  }
}

SDK (LangGraph / custom Python):

# before
tool.send_email(payload)

# after
guardyx.execute("send_email", payload)

Where Guardyx AI fits

Guardyx AI governs the runtime path agents use to act — not the paperwork around your company's compliance posture.

What it is

An infrastructure layer between agent frameworks and enterprise systems. Governs live tool execution with policy enforcement, approval workflows, observability, audit, and usage metering.

What it is not

Not a GRC or compliance workflow product. No questionnaires, no evidence collection, no trust-center automation. Guardyx AI starts where agents actually act — the tool invocation path.

vs. compliance tools

Vanta and Drata help govern the company's compliance posture.

Guardyx AI helps govern the agent's runtime behavior.

Get Early Access

Pilot Guardyx AI with your existing agents.

Guardyx is in active development. We're onboarding a small group of design partners to shape what ships — free during the beta, with hands-on onboarding.

Create an account