MCP API Tester

Mcp Api Tester is a lightweight MCP server that equips AI assistants with the ability to perform end‑to‑end automated testing of RESTful APIs directly from documentation. By exposing a set of high‑level tools, the server lets an LLM read OpenAPI or similar specs, discover available endpoints, generate realistic request payloads, execute those calls, and validate the responses—all without leaving the conversational context.

The core of the server revolves around three minimal‑viable tools. First, listAllAPIFromDocument parses the documentation and returns a concise catalog of endpoints, including names, URLs, and HTTP methods. This enables the LLM to browse the API surface and pick which parts to test. Second, getSingleAPIDetail dives deeper into a chosen endpoint, exposing its parameters, expected responses, status codes, and error definitions. With this information, the LLM can craft precise test cases that mirror real usage scenarios. Third, callNetHTTP actually sends an HTTP request and streams back the full response—status code, headers, and body. Together, these tools give an AI assistant a complete testing loop: discover → understand → execute.

Beyond the essentials, the server offers optional advanced utilities that elevate test quality and maintainability. Schema‑validation helpers compare live responses against documented JSON schemas, flagging missing or incorrectly typed fields. Parsing tools make it trivial for the LLM to extract specific values from responses, enabling conditional logic in tests. Logging and storage utilities let the assistant record each test run, capture failures, and retrieve historical results to spot regressions. For authenticated APIs, tools such as getAuthToken and environment‑reset helpers allow the LLM to manage credentials and test data cleanly.

In practice, this MCP server is invaluable for developers who want to prototype API tests in natural language or integrate automated checks into CI pipelines. A product owner can ask the assistant, “Run a smoke test on the user‑creation endpoint,” and the LLM will automatically list available endpoints, fetch the required parameters, execute a POST request, validate the response against the schema, and report success or failure—all while keeping the conversation context. The server’s modular design means it can be extended with custom validation logic or additional environment‑management tools, making it a versatile foundation for AI‑driven API quality assurance.