About Preloop

AI safety, control, and observability for teams that need speed and control

Preloop helps companies adopt AI agents in real production workflows without giving up control. As agents gain access to codebases, infrastructure, internal tools, business systems, and model budgets, teams need more than prompts and trust. They need policies, approvals, visibility, auditability, and budget controls.

That is what Preloop provides.

We build the control plane for AI agents so teams can move faster with OpenClaw, OpenCode, Claude Code, Codex CLI, Gemini CLI, Cursor, Windsurf, and other MCP-compatible tools, while still keeping humans in control of consequential actions.

Our mission

Our mission is to make AI automation safe, observable, and governable enough for real work.

We believe companies should not have to choose between:

locking agents down so tightly that they are no longer useful
giving agents broad access with no reliable controls

Preloop is designed to give teams a better middle ground: let low-risk actions run automatically, require approval for sensitive actions, keep model traffic and spend visible, and keep a complete audit trail of what agents do.

The problem we are solving

AI agents are becoming operational software, not just assistants.

They can now:

open pull requests and merge code
deploy services and change infrastructure
read and modify production data
call internal tools and business workflows
spend money through cloud, billing, and procurement systems

That creates a new control problem.

Traditional IAM, prompt rules, and manual reviews were not built for agentic workflows. Teams need controls that are designed for AI actions happening in real time, across tools, with enough context to decide what should run automatically and what should stop for review.

What Preloop does

Preloop gives engineering, platform, security, and operations teams a control plane for AI agents.

With Preloop, teams can:

onboard existing local agents with the preloop agents discover command
configure event driven flows to automate processes like code reviews, system operations, and more
define allow, deny, and approval policies for AI tool calls
require human sign-off for sensitive operations
route approvals to the right person or team
route model traffic through the Preloop Gateway for attribution and budget visibility
govern tool access through the Preloop Tool Firewall
inspect runtime sessions, usage, and spend with operator visibility
inspect every action with context, policy decisions, and outcomes
keep a durable audit trail for security reviews and compliance
define and enforce budgets per agent and per model

Preloop can be adopted without rebuilding the rest of your stack. For teams evaluating operational AI governance controls, see EU AI Act readiness with Preloop. We are available to help you get started automating new workflows or onboarding existing agents. Book a demo.

Who Preloop is for

Preloop is for teams that want AI agents to do real work, not just demos.

Engineering and platform teams

Ship faster while protecting production deploys, secrets, cloud resources, databases, and internal tooling.

Security and compliance teams

Gain visibility into what agents are doing, which policies are enforced, and who approved sensitive actions.

Founders and operators

Adopt AI automation with clearer accountability, lower operational risk, and less fear of expensive mistakes.

What we believe

AI agents need governance, not just access
Human approval should be easy for high-impact actions
Good controls should preserve speed, not kill it
Every important AI action should be visible and explainable
Model traffic and spend should be governed, not treated as a blind spot
Teams should be able to adopt AI safely without rewriting their infrastructure

Our team

Preloop is built by experienced founders and engineers who have worked across AI, infrastructure, robotics, product, and developer tooling. We care deeply about making powerful systems usable in the real world, especially when the cost of mistakes is high.

Founders

Dimitris Moraitis

Founder & CEO

Dimitris is a software engineer and entrepreneur focused on AI, infrastructure, DevOps, and product UX. He founded Mist.io, the open-source multicloud management platform, and served as CTO through its acquisition by Dell Technologies. After the acquisition, he worked at Dell as an Edge Software Architect. In 2025, Dimitris co-founded Preloop to drive enterprise adoption of AI agents while keeping humans in control of important decisions.

LinkedIn · GitHub

Dr. Yannis Chatzikonstantinou

Advisor of Product Strategy

Yannis is an AI and robotics engineer with deep expertise in evolutionary computation, machine learning, and product design. He holds a PhD from TU Delft and has worked across academia and industry in AI and robotics. He brings together AI research, software engineering, and product thinking to build systems that help teams adopt AI in practical, reliable ways.