The Safety Layer for AI Agents

Control what AI agents can do, and what requires human approval.

Sign In /login

Define policies for AI agents & tools. Allow, deny, or require approval for any action based on context. Full visibility and audit trail for every AI operation. Works with OpenClaw, Claude Code, Cursor, and any MCP-compatible agent.

grid Onboard or connect your AI agent in minutes Read the docs https://docs.preloop.ai Start with a simple CLI setup, or connect Preloop manually