Wassette: A Security-Oriented Runtime for WebAssembly Components
Summary
Wassette is a security-focused runtime developed by Microsoft, designed to execute WebAssembly Components via the MCP protocol. It provides a secure sandbox environment, making it easy to extend AI agents with new tools. Leveraging Wasmtime, Wassette offers browser-grade isolation for enhanced security, though it is currently in early development.
Repository Info
Introduction
Wassette is a security-oriented runtime developed by Microsoft, built to run WebAssembly Components through the MCP (Micro-Component Protocol). It aims to provide a robust and secure environment for integrating new functionalities, particularly for extending AI agents with various tools. While currently in early development, Wassette emphasizes security by design, utilizing the Wasmtime security sandbox for strong isolation.
Installation
Getting started with Wassette is straightforward. For Linux and macOS users, a convenient one-liner script is available:
curl -fsSL https://raw.githubusercontent.com/microsoft/wassette/main/install.sh | bash
For comprehensive installation instructions covering all platforms, including Windows, Homebrew, Nix, and Docker, please refer to the official Installation Guide.
Examples
Once Wassette is installed, the next step involves registering it with your chosen AI agent. The Quick Start guide provides a simple three-step process, and the MCP Clients guide offers detailed setup instructions for various agents like GitHub Copilot, Cursor, Claude Code, and Gemini CLI.
After configuration, you can begin loading WebAssembly components. For instance, to enable your agent to tell the time, instruct it to load a time component:
Please load the time component from oci://ghcr.io/microsoft/time-server-js:latest
With the time component loaded, you can then ask your agent for the current time:
What is the current time?
The agent will respond with the current time, fetched securely from the WebAssembly component running within Wassette's sandbox:
The current time July 31, 2025 at 10:30 AM UTC
This demonstrates a basic interaction, showcasing how Wassette allows agents to securely execute external functionalities.
Why Use Wassette?
Wassette offers several compelling advantages for developers and AI practitioners:
- Convenience: It simplifies the process of extending AI agents with new tools, allowing seamless integration without leaving the chat interface.
- Reusability: WebAssembly Components are inherently generic and reusable, meaning they are not tied to MCP-specific implementations and can be deployed across various contexts.
- Security: Built upon the robust Wasmtime security sandbox, Wassette provides browser-grade isolation for tools, ensuring a high level of security and preventing unauthorized access or malicious behavior.