home / mcp / mcp allure mcp server

MCP Allure MCP Server

Converts Allure test reports into LLM-friendly JSON for AI-powered analysis and documentation.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "crisschan-mcp-allure": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "mcp[cli]",
        "mcp",
        "run",
        "/Users/crisschan/workspace/pyspace/mcp-allure/mcp-allure-server.py"
      ]
    }
  }
}

MCP-Allure is an MCP server that converts Allure test reports into formats that are easy for language models to consume, enabling AI-assisted analysis, summaries, and debugging of test results.

How to use

You run MCP-Allure as an MCP server beside your existing tooling, then use an MCP client to interact with the server and access the transformed Allure data. The server exposes a tool named get_allure_report that reads an Allure report directory and returns structured JSON data. You can pass the Allure HTML report path, and you’ll receive a JSON payload that contains test suites, test cases, steps, and labels suitable for AI processing.

To work with the server, first start it using the provided MCP runtime command. Then, use your MCP client to request the get_allure_report tool, pointing it at your Allure report directory. The tool will return a JSON string that includes all pertinent details about suites, cases, and steps, ready for AI analysis.

How to install

Prerequisites: you need a runtime capable of loading MCP servers (the MCP runtime) and the uv execution tool used to launch servers via MCP configurations.

Create or update your MCP configuration with the following snippet. This defines a stdio MCP server named mcp_allure that runs through the uv runtime and starts the mcp_allure-server.py entry point.

{
  "mcpServers": {
    "mcp_allure": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "mcp[cli]",
        "mcp",
        "run",
        "/Users/crisschan/workspace/pyspace/mcp-allure/mcp-allure-server.py"
      ]
    }
  }
}

Configuration and notes

The server is configured to run with the MCP runtime via uv. You can see the exact startup command and arguments in the configuration block above. Ensure the path to mcp-allure-server.py accurately reflects your environment.

Additional details

Tool exposed by this server: get_allure_report reads an Allure report directory and returns JSON data representing test Suites, Test Cases, and Steps. Example input is a directory containing Allure HTML reports, and the output is a JSON payload suitable for downstream AI processing.

Available tools

get_allure_report

Reads an Allure report from a specified directory and returns a JSON string containing the report structure, including test suites, test cases, steps, labels, and attachments.