Guides

Task walkthroughs, each step shown with a real terminal recording or console screenshot.

Getting started

Enroll a device to the console

Mint a token, enroll a machine, and watch its first events land in the console.

Cover Codex, Cursor, Copilot and Antigravity

Hook every coding agent on the machine with one policy.

Coming soon

Cloak your secrets

Import a .env, let the agent use placeholders, and keep real values out of its context.

Policy

Roll out from observe to enforce

Watch first, pick the rules that block, and carve narrow exceptions.

Write a custom rule

Author a rule in YAML, test it locally, and push it to the fleet.

Coming soon

Egress allowlist and access requests

Limit where agents can send data and approve new domains on request.

Console

Investigate a session

Follow an agent session turn by turn, ask why a call was blocked, and share it.

Approve risky calls from Slack

Route ask decisions to a Slack channel and approve them in one click.

Find shadow AI

Inventory agents, MCP servers and keys nobody registered.

Teams, SSO and agent keys

Invite your team, connect your IdP, and issue keys for headless agents.

Advanced

Put every MCP server behind one gateway

One connector, every tool call policy-checked before it runs.

Govern any app with the LLM proxy

Swap a base URL to screen n8n workflows and Vercel AI SDK apps.

Turn on the semantic judge

Add an LLM verdict on top of the rules and read its reasoning.

Govern production framework agents

Run OpenAI Agents, LangGraph, CrewAI, Google ADK and Microsoft Agent Framework under one key and one policy.

Guard an SDK agent

Wrap an OpenAI Agents, LangChain or CrewAI agent in about ten lines.