Skip to main content

​
Get started

Runlayer separates day-to-day usage from setup and governance:
  • End users access org-approved tools, skills, and agents through the Runlayer Plugin or Agents — they see a curated surface, not the setup behind it. Users without an admin role get the default User role and can only use MCPs they’ve been granted access to.
  • Builders and developers create connectors, skills, and agents from the dashboard, and can deploy custom MCP servers with Runlayer Deploy.
  • Admins assign roles, define policies, and monitor activity with sessions and audit logs.
See Roles & Permissions for the full access model, including how platform-feature access (FGA) and MCP tool invocation (PBAC) are enforced independently.

​
Enable

Connect. Build. Run AI at scale, wherever your workflow lives.

Connectors

Connect AI to your tools through one gateway

Skills

Create, share, and install skills across your teams

Agents

Build and run agents from where you work

Artifacts

Create, find, and share reports, interactive pages, and files

​
Control

Gain full visibility and control over agent activity with real-time monitoring, policy management, and advanced threat detection.

Shadow AI

Discover and govern unsanctioned AI, MCPs and skills

Runtime Monitoring

See, alert, and block any agent session

Policies

Define fine-grained, context-aware access controls

​
Deploy Runlayer

Runlayer runs as a hosted deployment or self-hosted in your own AWS account, integrating with existing Terraform and Kubernetes workflows.

Deployment Overview

Compare deployment methods and pick the right fit

ECS + Terraform

Deploy on AWS ECS Fargate with Terraform

EKS & Kubernetes

Runlayer Operator on EKS or GKE (customer-managed cluster)

​
Build on Runlayer

Custom MCP Servers

Build and deploy your own MCP servers on Runlayer Deploy

Hooks SDK

Add hook enforcement and telemetry to custom agents

Platform API

Automate Runlayer programmatically from your own tooling