Webpage Timestamps

You can extract creation, modification, and publication timestamps from webpages using this MCP server. It gathers data from multiple sources, assigns confidence levels, and supports batch processing, making it practical for web archives, SEO analysis, and research on content lifecycles.

How to use

You run this MCP server through an MCP client to extract timestamps for one or more webpages. Start the server with your preferred method, then send requests for URLs you want analyzed. The server returns the timestamps it finds, the sources it used, and a confidence rating so you can trust or verify the results.

Practical usage patterns include running a single URL fetch or batching several URLs to process together. You can tune timeout, user agent, and redirect behavior to optimize results for different websites or data requirements.

If you are using Claude Desktop, you configure the MCP server with a dedicated entry under mcpServers. The server is invoked via npx and the package name mcp-webpage-timestamps, so your client will start the server in a lightweight, publish-ready way.

How to install

Prerequisites you need before installing: Node.js 18.0.0 or higher, and npm or yarn for package management.

Quick install (global CLI): you install the MCP server package globally, which makes the mcp-webpage-timestamps command available from any location.

Usage with npx lets you run the MCP server directly without a global install.

Development install steps are for building and testing from source when you want to contribute or customize behavior.

Configuration and server endpoints

To use this MCP server, you typically configure your MCP client to point at the server using an MCP entry. The examples below show two common configurations for Claude Desktop and Cline clients.

Claude Desktop configuration example shows how to declare the server in claude_desktop_config.json with the command npx and the argument mcp-webpage-timestamps. An empty env object is shown for completeness.

Cline configuration example mirrors the Claude setup but without the explicit env section.

These configurations instruct the MCP client to launch the server process when needed and to route timestamp extraction requests to it.

Examples of starting and using the server

Start the server directly if you have a local setup that supports npm scripts. You can initiate the server in development mode or as a production-ready process.

For local development, you typically run npm start or npm run dev to start the server and enable hot reloading or production-like behavior, depending on your environment.

Tools and endpoints overview

The server exposes two main tools you use from an MCP client: extract_timestamps for a single URL and batch_extract_timestamps for multiple URLs. Each tool returns a structured JSON response containing the url, timestamp fields, sources, confidence, and any errors.

Notes on usage and reliability

The server aggregates data from various sources such as HTML meta tags, HTTP headers, JSON-LD, microdata, OpenGraph, Twitter cards, and heuristic analysis. It reports a confidence level for each extracted timestamp to help you assess reliability.

Troubleshooting

If requests time out or return HTTP errors, adjust the timeout and redirect handling in your client configuration. Ensure URLs are reachable and that the target pages expose timestamp metadata through at least one supported source.

Development scripts and quality tooling

Development workflows include hot reloading, building, running tests, linting, and formatting. Use npm run dev for development with hot reload, npm run build to build the project, and npm test to run tests. Code quality is maintained with TypeScript, ESLint, Prettier, and Jest, with a goal of high test coverage.

Available tools

extract_timestamps

Extract timestamps from a single webpage, returning the URL, createdAt, modifiedAt, publishedAt, sources, confidence, and errors.

batch_extract_timestamps

Extract timestamps from multiple webpages in a batch, returning an array of similar timestamp results for each URL.