Playwright-BDD-POM-MCP Server 🚀

A powerful Model Context Protocol (MCP) server designed to automate Gherkin-based test generation, execution, and self-healing for Playwright-BDD projects using the Page Object Model (POM) pattern.

Features

Project Bootstrapping: Scaffolds full Playwright-BDD environments from scratch (setup_project).
Smart Analysis: Scans your codebase to reuse existing Page Objects and step definitions (analyze_codebase).
Gherkin Generation: Generates high-quality .feature and .ts files based on your specific patterns.
Self-Healing: Automatically detects failures, inspects the DOM, and fixes locators/scripts up to 3 times (validate_and_write).
Flexible Config: Custom tags, directory structures, and wait strategies via mcp-config.json.
Multi-User Store: Environment-specific JSON credential management (users.{env}.json) with a typed getUser() helper.

📚 Documentation

For detailed guides and advanced usage, refer to the following resources:

User Guide: Learn about "Power Prompts", API fixtures, and TypeScript DTO generation.
Docker & Deployment: Comprehensive guide for local, remote, and AWS containerization.
Project Evolution: A technical look at the project's architecture and development milestones.
Implementation Plan: Detailed roadmap of recent and upcoming feature phases.

Quickstart

1. Installation

# Install from Git (Recommended for latest features)
npm install -g github:Rohit5688/playwright-bdd-mcp

# OR Clone and build locally
git clone https://github.com/Rohit5688/playwright-bdd-mcp.git
cd playwright-bdd-mcp

# Install and build
npm install
npm run build

2. Configure your MCP Client (e.g. Claude Desktop)

Add the server to your claude_desktop_config.json:

{
  "mcpServers": {
    "playwright-bdd-automation": {
      "command": "node",
      "args": ["/absolute/path/to/playwright-bdd-pom-mcp/dist/index.js"]
    }
  }
}

3. Initialize a New Project

Ask your AI assistant:

"Use the setup_project tool to initialize a new Playwright-BDD automation suite in C:/my-tests."

Tool Reference

| Tool | Action | Summary | |---|---|---| | analyze_codebase | Context | Scans project for existing POMs, steps, and mcp-config.json rules. | | generate_gherkin_pom_test_suite | AI | Returns a rigid generation instruction for high-quality test code. | | validate_and_write | Write + Test | Writes the generated code, runs it, and auto-heals failures. | | run_playwright_test | Run | Executes tests with optional tag filters (e.g., @smoke). | | inspect_page_dom | Debug | Returns a semantic Accessibility Tree (AOM) for locator extraction. | | manage_config | Config | Reads/writes/scaffolds mcp-config.json team preferences. | | manage_users | Security | Manages environment-specific user roles and credentials. | | summarize_suite | Report | Provides a plain-English status report of the entire test suite. |

Governance & Configuration

`mcp-config.json`

Located in your project root, this file controls global behavior:

{
  "tags": ["@smoke", "@regression", "@p1"],
  "waitStrategy": "networkidle",
  "selfHealMaxRetries": 3,
  "currentEnvironment": "staging"
}

Multi-User Credentials

Credentials are stored in test-data/users.{env}.json (e.g. users.staging.json).

These files are git-ignored for security.
Use manage_users { action: "scaffold" } to generate templates.

Access them in your Page Objects via the generated user-helper.ts:

import { getUser } from '../test-data/user-helper.js';
const { username, password } = getUser('admin');

Technical Details

Base image: mcr.microsoft.com/playwright
Transport: Supports both stdio (default) and sse (via --port).
Dependencies: @modelcontextprotocol/sdk, playwright-bdd, @playwright/test.

License

ISC

MCP Servers