MCP Toolhouse

MCP Toolhouse is an MCP server that integrates external tools from the Toolhouse platform into Model Context Protocol environments. It connects AI assistants directly to curated Toolhouse bundles, allowing models to invoke external workflows, run developer integrations, and manage autonomous agent actions. Developers, engineers, and AI practitioners use this server to expand the practical utility of their local or cloud-based large language models without writing custom API adapters for every separate service. By bridging the Toolhouse catalog with standard MCP clients, MCP Toolhouse enables agents to query third-party platforms, execute specialized scripts, and trigger automated tasks securely. The server runs via standard Python execution tooling, consuming API keys and bundle configurations directly through environment variables. This approach minimizes setup complexity while standardizing external tool execution inside supported client ecosystems like Claude Desktop.

Category: AI & LLM Tooling

Tags: agent-tools, automation, integrations, toolhouse

Visit MCP Toolhouse

How to install and configure MCP Toolhouse

Follow these steps to set up and register MCP Toolhouse: 1. Automatically install for Claude Desktop using Smithery by running: npx -y @smithery/cli install @SecretiveShell/MCP-toolhouse --client claude 2. Alternatively, configure manual installation by updating your MCP configuration file (such as claude_desktop_config.json) with the following block under mcpServers: json { "mcpServers": { "MCP-Toolhouse": { "command": "uv", "args": ["run", "mcp-toolhouse"], "env": { "TOOLHOUSE_API_KEY": "<YOUR_TOOLHOUSE_API_KEY>", "TOOLHOUSE_BUNDLE": "<YOUR_TOOLHOUSE_BUNDLE_NAME>", "PYTHONUTF8": "1" } } } } 3. Note that setting the PYTHONUTF8 environment variable to 1 is required for some tools to function properly on Windows operating systems. 4. Restart your MCP client to load the server and access your Toolhouse tools.

What you can do with MCP Toolhouse

• Exposing pre-configured Toolhouse tool bundles to LLMs in Claude Desktop for expanded agent utility. * Automating multi-step developer workflows by executing cloud-hosted actions through standardized MCP protocol commands. * Managing external third-party API interactions safely from local AI assistant environments without extra manual boilerplate. * Running cross-platform agent integrations on Windows systems with proper UTF-8 character encoding support enabled.

Key facts

• https://github.com/SecretiveShell/MCP-toolhouse
• AI & LLM Tooling, Developer Tools & Code Intelligence
• agent-tools, automation, integrations, toolhouse

Part of MCP Servers

What is MCP Toolhouse?

MCP Toolhouse is a Model Context Protocol server that connects AI clients to the Toolhouse platform. It allows supported language models to dynamically call external tools and integrations bundled inside Toolhouse.

How do I install MCP Toolhouse?

You can install it automatically using Smithery by running the npx command provided in the documentation. Alternatively, configure your MCP client manually to run mcp-toolhouse using uv, specifying your Toolhouse API key and bundle name in the environment settings.

Which MCP clients work with MCP Toolhouse?

MCP Toolhouse is confirmed to support Claude Desktop via automatic Smithery installation or manual configuration. It works with any MCP-compliant client capable of launching Python processes through uv and passing custom environment variables.

Why is the PYTHONUTF8 environment variable needed?

Setting PYTHONUTF8 to 1 is specified in the manual configuration to ensure that certain Toolhouse tools function correctly when operating on Windows systems, preventing character encoding conflicts.