Runlayer

Runlayer is an enterprise MCP (Model Context Protocol) gateway and AI control plane that lets security and IT teams govern how AI agents connect to internal tools and data. It is built for CISOs, CIOs and platform teams who want employees to build agents freely while keeping identity, permissions, audit and spend under central control.

Runlayer is a model-neutral control plane for enterprise AI agents, built around the Model Context Protocol (MCP) — the open specification that defines how AI agents plug into enterprise systems and data. The platform combines agent enablement with security and governance in one system: it discovers 'Shadow AI' (unmanaged agents and integrations already running inside the company), brokers agent access to more than 18,000 pre-built MCP connectors through a central gateway with a catalog and approval workflows, enforces identity-scoped permissions by integrating with providers such as Okta and Entra, scans tool calls in real time for prompt injection and sensitive-data leakage via its Guard capability, and attributes AI spend across teams. It supports the AI clients enterprises actually use — Claude, Cursor, ChatGPT, Codex, vertical AI apps, independent agents and platforms like Salesforce Agentforce. Founded in 2025 by three-time founder Andrew Berman with Tal Peretz and Vitor Balocco (all previously at Zapier), Runlayer launched from stealth in November 2025 with an $11M seed and raised a $30M Series A on June 24, 2026 co-led by Felicis and Khosla Ventures, bringing total funding to $42M. Early adopters include Gusto, Instacart, dbt Labs, Opendoor, Decagon, Lemonade, AngelList and PagerDuty.

Category

Governance & Security

Pricing

Contact for pricing

Target Market

CISOs, CIOs, CTOs, Platform Engineers, Enterprise Developers, IT Security Teams

Founded

2025

Customers

Eight unicorns plus several public companies, including Gusto, Instacart, dbt Labs, Opendoor, Decagon, Lemonade, AngelList and PagerDuty

• ✓
  Enterprise MCP gateway

  A central gateway that brokers every AI client's MCP requests to internal tools and data, with a catalog of 18,000+ pre-built connectors and approval workflows.

• ✓
  Shadow AI discovery

  Surfaces unmanaged AI agents, MCP servers and integrations already running across the company so security teams can bring them under policy.

• ✓
  Identity-scoped permissions

  Integrates with identity providers such as Okta and Entra so an agent inherits only the permissions of the user it acts for.

• ✓
  Guard real-time scanning

  Inspects tool calls and returned data in real time for prompt injection and sensitive-data leakage before they reach the model.

• ✓
  Agent builder

  An on-demand builder that lets employees compose agents from reusable skills, internal APIs and approved connectors.

• ✓
  AI spend attribution

  Monitors AI usage and cost and attributes it back to teams, agents and tools for chargeback and budgeting.

• •
  Safe employee-built agents

  Let engineering, ops and business teams build their own agents against a pre-vetted connector catalog instead of wiring unmanaged MCP servers themselves.

• •
  Agent security and audit

  Give security teams a single audit trail and real-time threat detection across every agentic action taken in the company.

• •
  Shadow AI governance

  Find and govern the AI clients and MCP integrations employees have already connected to production systems.

Runlayer does not publish pricing; the product is sold as an enterprise engagement.

This page was written from 3 sources, 1 on domains other than runlayer.com.

1. 1.runlayer.com — runlayer.comvendor
2. 2.runlayer.com — series A 30m fundraise felicis khoslavendor
3. 3.techcrunch.com — mcp ai agent security startup runlayer launches with 8 unico