logo
Free, unlimited AI code reviews that run on commit
git-lrc git-lrc GitHub Install Now We'd appreciate a star git-lrc - Free, unlimited AI code reviews that run on commit | Product Hunt git-lrc - Free, unlimited AI code reviews that run on commit | Product Hunt

playwright-toolkit-connector

A specialized MCP service enabling sophisticated web interaction via Playwright, featuring integrated persistent knowledge storage with flexible summarization capabilities.

Author

playwright-toolkit-connector logo

blackwhite084

Apache License 2.0

Quick Info

GitHub GitHub Stars 153
NPM Weekly Downloads 0
Tools 1
Last Updated 2026-02-19

Tags

automationscrapingwebautomation webweb automationbrowser automation

Playwright Automation & Knowledge Server

\A Bridge for automated web control leveraging Playwright\

Playwright Server MCP server

Operational Components

Data Repository

This server maintains a rudimentary persistent data mechanism for notes: - Utilizes a custom note:// Uniform Resource Identifier scheme for individual data item retrieval. - Each stored note entity comprises a unique identifier, a descriptive summary, and content transmitted with a text/plain MIME type.

Interaction Routines (Prompts)

The service exposes one primary processing endpoint: - summarize-notes: Triggers the aggregation and condensation of all presently stored knowledge entries. - Accepts an optional style parameter to modulate the output verbosity (e.g., 'brief' or 'detailed'). - Dynamically constructs a comprehensive prompt that incorporates all notes alongside the specified level of detail.

Utility Functions (Tools)

The server exposes the following granular automation capabilities: - playwright_navigate: Directs the underlying browser instance to a specified Uniform Resource Locator (URL). A new browsing context is automatically initiated if none is currently active. - Mandatory parameter: url (string). - playwright_screenshot: Captures the visual state of the current viewport or a targeted element. - Mandatory parameter: name (string) for the output file designation. - Optional parameter: selector (string) identifying a specific Document Object Model (DOM) element via CSS query. Omitting this results in a full-page capture. - playwright_click: Simulates a user mouse click event upon an element identified by its CSS selector. - Mandatory parameter: selector (string). - playwright_fill: Inputs specified data into a designated interactive form field. - Mandatory parameter: selector (string) locating the input element. - Mandatory parameter: value (string) representing the data payload to insert. - playwright_evaluate: Executes arbitrary JavaScript code within the isolated browser execution environment. - Mandatory parameter: script (string) containing the JavaScript payload. - playwright_click_text: Locates and interacts (clicks) with an element based on its visible textual content. - Mandatory parameter: text (string) matching the target element's label. - playwright_get_text_content: Retrieves the textual data from all currently rendered, visible elements on the page. - playwright_get_html_content: Fetches the complete HyperText Markup Language structure. - Optional parameter: selector (string) to retrieve the inner HTML of a specific node rather than the entire document body.

System Configuration

[TODO: Placeholder for environment-specific setup parameters]

Initial Deployment Guide

Installation Procedure

End-User Desktop Integration (Claude)

For macOS environments: ~/Library/Application\ Support/Claude/claude_desktop_config.json For Windows environments: %APPDATA%/Claude/claude_desktop_config.json

Configuration for Local/Unreleased Instances "mcpServers": { "playwright-server": { "command": "uv", "args": [ "--directory", "C:\\Users\\xxxxx\\Documents\\project\\python\\mcp\\playwright-server", "run", "playwright-server" ] } }
Configuration for Officially Distributed Instances "mcpServers": { "playwright-server": { "command": "uvx", "args": [ "playwright-server" ] } }

Development Workflow

Artifact Generation and Distribution

Steps required to package and deploy the service:

  1. Synchronize project dependencies and update the dependency lock file: bash uv sync

  2. Compile the distribution artifacts (Source and Wheel formats): bash uv build

  3. Push packages to the Python Package Index (PyPI): bash uv publish

Note: Authentication credentials for PyPI must be provided, either via command-line flags (--token, --username/--password) or environment variables (e.g., UV_PUBLISH_TOKEN).

Troubleshooting and Inspection

Debugging interactions over standard input/output (stdio) can be complex. We highly recommend utilizing the official MCP Inspector utility for a superior debugging experience.

You may initialize the Inspector via Node Package Manager (npm):

bash npx @modelcontextprotocol/inspector uv --directory C:\Users\YUNYING\Documents\project\python\mcp\playwright-server run playwright-server

Once active, the Inspector will output a local URL. Accessing this URL in a standard web browser initiates the debugging session.

See Also

`