Overview
What is Riza MCP Server?
Riza MCP Server wraps the Riza API—an isolated code interpreter for LLM-generated code—and exposes its endpoints as individual tools. It is intended for developers who want to let their language models write and run code securely.
How to use Riza MCP Server?
Configure the server with your MCP client (e.g., Claude Desktop) using the npx command @riza-io/riza-mcp and set the environment variable RIZA_API_KEY to your API key from the Riza Dashboard. Once running, your LLM can call any of the server’s tools directly.
Key features of Riza MCP Server
- Wraps Riza API as MCP tools for LLMs.
- Execute arbitrary code safely without saving.
- Create, list, fetch, edit, and execute saved tools.
- Uses Riza’s isolated code interpreter for security.
- Simple setup with a single API key.
Use cases of Riza MCP Server
- Let an LLM write and run Python/JavaScript code on demand.
- Build and maintain a library of reusable code tools.
- Enable your LLM to edit and refine its own saved tools.
- Run untrusted code in a sandboxed environment.
FAQ from Riza MCP Server
What does Riza MCP Server do that a plain code interpreter doesn’t?
It provides a set of MCP tools that give your LLM fine-grained control over code execution and tool management (create, read, update, delete) without exposing the raw API.
What runtime or dependencies are required?
Node.js and npm are required to run the npx command. No additional dependencies beyond a Riza API key.
Where does the code execution happen?
All code runs on Riza’s isolated interpreter infrastructure, not on the client machine.
What transports or authentication does it use?
The server communicates with the Riza API using your provided API key. The MCP transport is standard stdio via the command configured in the client (e.g., npx).
Are there any known limits on code execution?
The README does not detail limits; refer to the Riza API documentation for execution time, memory, or file system restrictions.
