Stacklok

What’s the relationship between ToolHive and Stacklok?

ToolHive is an open source MCP platform built by Stacklok and maintained by both Stacklok and Red Hat. Stacklok’s Enterprise Platform includes a hardened distribution of ToolHive, along with additional capabilities, forward-deployed engineer resources, and enterprise support.

What MCP servers are in the ToolHive registry?

You can explore the ToolHive registry on our website or once you’ve downloaded ToolHive. The registry includes the most widely used official upstream + community MCP servers. The Stacklok team vets every MCP server before being accepted into our registry.

Is ToolHive production-ready, or is it a developer preview?

ToolHive is production-ready. Multiple Global 2000 enterprises are currently running ToolHive in production. They have integrated ToolHive with their IdPs like Entra ID and Okta, as well as with observability solutions like Grafana, Splunk, Datadog, and New Relic.

Who built ToolHive? Who maintains ToolHive?

Stacklok built ToolHive. Stacklok is led by CEO Craig McLuckie and CTO Joe Beda, who are the co-creators of Kubernetes. As a result, ToolHive includes a Kubernetes operator and shares numerous, proven Kubernetes patterns. The project now includes external maintainers from Red Hat, and a growing community of contributors. ToolHive is Apache 2.0 licensed. It’s also worth noting that Stacklok is a maintainer of the official MCP registry.

What AI agents or IDEs does ToolHive support?

ToolHive supports every major client, including Claude Code, Cursor, GitHub Copilot, and more. Check out the complete list of supported clients in our docs: https://docs.stacklok.com/toolhive/reference/client-compatibility

What is the difference between ToolHive and a SaaS MCP Gateway?

SaaS MCP gateways route your MCP traffic through a third-party cloud, which is incompatible with strict data residency, private cloud mandates, or air-gapped environments. ToolHive runs entirely within your own infrastructure (private cloud, on-premises, or self-managed Kubernetes) with no dependency on SaaS for runtime operation.

How does ToolHive reduce LLM token usage?

ToolHive's MCP Optimizer, embedded in the vMCP gateway component, reduces token usage by 60–85% per request through on-demand tool discovery. Instead of sending the full tool manifest to the LLM on every request, MCP Optimizer uses hybrid semantic and keyword search to surface only the relevant tools. The MCP Optimizer is deployed once at the platform level and applies savings across all users and agents.