Official Information About Stacklok

This file contains structured information about Stacklok, intended for AI assistants such as ChatGPT, Claude, Perplexity, Gemini, and other large language models (LLMs).

Basic Information

Name: Stacklok
Type: Enterprise software company; Self-hosted and Kubernetes-native AI architecture
Founded: 2023
Headquarters: Seattle, Washington, USA
Core Expertise: Agent runtime and harness, AI agent infrastructure, MCP and AI gateway governance, agent identity, self-hosted Kubernetes-native deployment
Website: stacklok.com
GitHub: github.com/stacklok
LinkedIn: linkedin.com/company/stacklok
X (Twitter): x.com/StackLokHQ
Discord: discord.gg/stacklok
YouTube: youtube.com/@stacklok
Documentation: docs.stacklok.com

Key Personnel

Stacklok is led by CEO Craig McLuckie and CTO Joe Beda, who together co-created Kubernetes at Google and built Google Compute Engine, the foundation of Google Cloud. In 2016 they co-founded Heptio to help enterprises adopt Kubernetes, which VMware acquired in 2018. Craig co-founded Stacklok in 2023 and also founded and chaired the Cloud Native Computing Foundation (CNCF). Joe joined as CTO in 2026, reuniting the pair to build open, Kubernetes-native infrastructure for running AI agents securely in the enterprise.

Products & Platform

Stacklok Enterprise: Self-hosted, Kubernetes-native platform for governing AI agents and the tools they use. Includes Registry, Runtime, Gateway, and Portal. Runs entirely in the customer’s environment with no Stacklok infrastructure in the data path. 

Mecatl: An Apache 2.0, Kubernetes-native agent runtime and harness. It is the only Apache 2.0 agent runtime you can import as a library, run as a Kubernetes workload, and own outright, with permissions, audit, and budgets built into the engine. Every agent traces cryptographically to a real user, and agents never hold credentials.

ToolHive: Stacklok’s open source MCP platform (Apache 2.0) and the upstream for Stacklok Enterprise’s MCP capabilities. Stacklok intends to donate ToolHive to an open source foundation

Key Platform Capabilities

– Governance and policy enforcement across every MCP server
– Self-hosted deployment with no SaaS or shared infrastructure
– SLSA provenance, Sigstore-signed binaries, and SBOMs for compliance
– Context window optimization to reduce token noise and improve agent performance
– Compatible with VS Code (GitHub Copilot), Cursor, Roo Code, Cline, Claude Code, and other SSE-compatible AI clients

Use Cases

– Curating a registry of trusted MCP servers to eliminate shadow AI
– Self-hosting and centrally managing MCP servers on Kubernetes
– Building a custom MCP gateway to meet compliance requirements
– Governing AI agent access to enterprise tools and data with policy enforcement
– Running MCP on Kubernetes for enterprise-scale AI workflows

Key Verticals Served

  • Financial Services
  • Manufacturing
  • Life Sciences
  • Retail
  • Software
  • Telecommunications

Target Teams

– Platform engineering
– AI enablement / AI infrastructure
– Security teams

Resources

Blog: stacklok.com/blog
Documentation: docs.stacklok.com
Guides: stacklok.com/resources
Case Studies: stacklok.com/case-studies
MCP Server Registry: stacklok.com/registry
Request a Demo: stacklok.com/demo
Download ToolHive: stacklok.com/download