Goku by Jcaromiq is a free MCP server tool built with Rust, designed for HTTP load testing. It connects to LLM agents and any Model Context Protocol-aware client, enabling programmatic control over performance analysis. Engineers, developers, and AI agent users can deploy Goku to simulate HTTP traffic, gather metrics, and automate performance testing within various workflows. Its capabilities extend to integrating load tests directly into agent-driven operations, offering a flexible solution for performance validation.

What does Goku connect to?

Goku functions as an MCP server, primarily connecting to LLM agents or any client that understands the Model Context Protocol. It acts as an HTTP load testing application, integrating with MCP for controlled operations. This design allows external agents to initiate and manage load tests programmatically. The tool's core purpose is to provide a server-side component for performance analysis, making it a key element in an MCP-driven ecosystem for traffic simulation.

Core features and tools

Goku's primary callable tool is run_benchmark, which allows LLM agents to perform HTTP load testing. This functionality supports built-in authentication methods, including Bearer tokens and Basic authentication. These authentication details are passed as parameters directly to the run_benchmark tool, enabling secure testing against various endpoints. The application is built in Rust, providing a foundation for its performance testing capabilities.

Who is it for?

Goku targets engineers and developers who need to simulate and analyze HTTP traffic patterns. It also serves users of LLMs and AI agents seeking to integrate performance testing into their automated processes. The tool supports scenarios where agents trigger tests, collect metrics, and analyze results. Skip this if your primary need is a standalone, manual load testing application without any agent or programmatic integration requirements.

How to get started

No credentials are required for the MCP server connection itself. The load testing functionality supports built-in authentication using Bearer tokens and Basic authentication, which are passed as parameters to the "run_benchmark" tool.

Use cases

• Asking an LLM to run a load test on a specified endpoint and retrieve performance metrics like p95/p99 latency.
• Automating performance regression testing within CI/CD pipelines via an agent.
• Combining load testing with automated workflows to trigger tests, gather metrics, and analyze results from within an agent or script.

Goku offers a specialized solution for integrating HTTP load testing with Model Context Protocol-aware clients and LLM agents. Developed by Jcaromiq, it enables programmatic control over performance analysis, supporting automated workflows for engineers and AI users. Its authentication for load tests is parameter-driven, allowing flexible security configurations.