ToolHive is an enterprise-grade open-source platform for running and managing Model Context Protocol (MCP) servers. Its core goal is to run any MCP server securely and instantly anywhere. ToolHive runs each MCP server in an isolated container, can enforce identity and access policies per request after configuring authentication sources, and provides platform teams with the observability needed to put MCP into production.
Key Capabilities
ToolHive uses a modular architecture. Its main components include the Gateway, Registry Server, and Runtime, accessible through the ToolHive UI and CLI.
- Gateway: Defines dedicated endpoints for teams to access tools securely and efficiently; can orchestrate multiple tools into a virtual MCP with a deterministic workflow engine; defines access policies and network endpoints; centrally controls security policies, authentication, authorization, auditing, and more; integrates with IdPs to implement SSO (compatible with OIDC/OAuth); customizes and filters tools and descriptions to improve performance and reduce token usage; connects local clients such as Claude Desktop, Cursor, VS Code, and VS Code Server.
- Registry Server: Curates a trusted server catalog for teams to quickly discover and deploy; integrates with the official MCP registry; adds custom MCP servers; groups servers by role or use case; manages the registry through an API-driven interface (or embeds it into existing workflows); verifies provenance and signs servers using built-in security controls; provides preset configurations and permissions to simplify the user experience.
- Runtime: Deploys, runs, and manages MCP servers locally or in Kubernetes clusters with security guardrails; deploys in the cloud through Kubernetes for enterprise-grade scalability; runs locally through Docker or Podman; securely proxies remote MCP servers for unified management; provides a Kubernetes Operator for cluster and resource management; leverages OpenTelemetry and Prometheus for monitoring and audit logging.
Usage
ToolHive provides multiple interfaces to suit different workflows:
- ToolHive UI: A cross-platform desktop experience, suitable for individual developers to get started quickly.
- ToolHive CLI: terminal workflows and automation, supporting discovery, configuration, and running of MCP servers, and automatically connecting managed MCP servers to compatible AI clients.
- Kubernetes Operator: teams and organizations use familiar Kubernetes workflows to centrally manage MCP servers and registries.
The quick start guide includes Quickstart and How-to guides for the desktop app, CLI, and Kubernetes Operator.
Use cases
- Individual developers: get started in minutes through the desktop UI or CLI, run any MCP server from a container image, or dynamically build from common package managers; manage encryption keys and network isolation; use built-in tools (such as the official MCP Inspector) to test and validate MCP servers; use MCP Optimizer to optimize token usage and tool execution.
- Teams and organizations: use Kubernetes workflows to centrally manage MCP servers and registries; define components such as MCP servers and registries through CRDs; container-based isolation and multi-namespace support for secure execution; automate service creation and discovery, with ingress integration for secure access; enterprise-grade security and observability, including OIDC/OAuth SSO, secure token exchange, audit logs, OpenTelemetry, and Prometheus metrics; hybrid registry servers that can curate from upstream registries, dynamically register local MCP servers, or proxy trusted remote services.
- Enterprises: integrate MCP servers into custom internal tools, agent workflows, or chat interfaces, using the same runtime and access controls.
Limitations and notes
ToolHive is licensed under Apache 2.0. The project welcomes community contributions and provides contribution guidelines, developer guides, and architecture documentation. Specific pricing, costs, platform support details, or performance data are not currently provided in the materials.
