Fetch FAQs

Question 1

Does Fetch allow for raw content retrieval without markdown conversion?

Accepted Answer

Absolutely. While its primary function is HTML to markdown conversion, Fetch provides a `raw` argument option. Setting this to `true` will retrieve the web page content in its original format without any markdown processing.

Question 2

What customization options are available for web fetching with Fetch?

Accepted Answer

Fetch offers extensive customization. Users can configure a custom `user-agent`, set up a proxy using `--proxy-url`, and control obedience to `robots.txt` files. These options provide flexibility for various web scraping and data collection scenarios.

Question 3

How can I install and integrate Fetch into my LLM environment?

Accepted Answer

Fetch can be easily installed via `uv` (recommended) or `pip`, and also supports Docker deployment. It integrates seamlessly with environments like Claude.app and VS Code, offering simple JSON configuration options to get started quickly.

Question 4

What is Fetch and how does it benefit LLMs?

Accepted Answer

Fetch is a Model Context Protocol (MCP) server designed to help Large Language Models (LLMs) retrieve and process web page content. It simplifies web data by converting complex HTML into easily consumable markdown format, enabling LLMs to understand and utilize web information more effectively.

Question 5

Can Fetch retrieve only parts of a webpage or truncate long responses?

Accepted Answer

Yes, Fetch supports advanced content handling. You can specify a `max_length` to truncate responses to a manageable size, and use the `start_index` argument to read web pages in chunks, allowing LLMs to efficiently process lengthy content incrementally.

Fetch

概要

主な機能

ユースケース