If you’re developing on a PC or just prefer a familiar environment, ToolHive makes installing and running MCP servers smooth, secure, and effortless on Windows.
What Is ToolHive?
ToolHive is an open-source platform that streamlines the deployment and management of Model Context Protocol (MCP) servers. It abstracts away the complexity of setup, containerization, runtime versioning, and security, letting you focus on innovation, not configuration.
ToolHive is available in three modes to meet different needs:
- Desktop app (ToolHive UI) for any developers on Windows, macOS, or Linux.
- CLI tool (thv) for terminal-first workflows, automation, and scripting.
- Kubernetes Operator for deploying and maintaining your remote MCP servers.
Running ToolHive on Windows
When we built-in Windows support, our goal was to keep things as simple as possible for the user. ToolHive includes:
- A native Windows installer (ToolHive.Setup.exe) for seamless setup.
- A desktop app that runs MCP servers is in isolated containers (via Docker Desktop or Podman), just like on other platforms.
- A UI that lets you browse, install, and manage MCP servers, no command line needed.
- Effortless setup for clients like GitHub Copilot, Cursor, Cline, Roo Code, and Claude Code, with automatic configuration.
Spinning Up an MCP Server Locally on Windows
Here’s a quick step-by-step walkthrough:
-
Download & Install ToolHive for Windows
Go to https://toolhive.dev/, click Download now to get the Windows desktop installer (
ToolHive.Setup.exe), then open it to begin installation.To run MCP servers, you’ll also need a container engine such as Docker Desktop, Podman, or Rancher Desktop (with dockerd).
-
Open and Launch the Desktop App (ToolHive UI)
ToolHive provides a clean UI with tabs like Installed, Registry, Clients, and Secrets to guide your workflow.
-
Browse and Install an MCP Server
Go to the Registry tab and select a verified MCP server, like fetch, which retrieves web content. Click Install Server, and ToolHive handles containerization, proxy setup (SSE-based), and launches the server.
-
Connect Your AI Client
In the Clients tab, toggle integration for your preferred editor or agent. ToolHive auto-configures it so your AI client can talk to the MCP server with zero fuss.
-
Use the MCP Server from Your AI Client
For example, ask GitHub Copilot to “fetch and summarize content from https://toolhive.dev/.” ToolHive forwards the request through the container to the MCP server, and your agent handles the response.
Why ToolHive Stands Out
Security by Default
Each MCP server runs in an isolated container with fine-grained permission control (filesystem, network, etc.). ToolHive also supports secret management through encrypted storage, steering clear of plaintext tokens. In addition, you can enable network isolation configuration for any server, giving you even tighter control over how and when it can access external resources.
Unified Experience Across Platforms
With the Windows installer, ToolHive maintains the same intuitive UI and workflows found on macOS and Linux. No need for platform-specific hacks or command-line heroes.
Flexible Developer Tools Compatibility
ToolHive auto-configures popular clients (Copilot, Cursor, etc.), but it also supports custom setups via SSE-based or manual configuration.
Runs Anything, Anywhere
Whether it’s curated MCP registry entries, custom Docker images, or even source packages (npx://, uvx://) that ToolHive builds into containers on-the-fly, ToolHive makes running MCP servers frictionless.
Top comments (0)