Skip to content
AI Dev Toolkit.
Esc
  • AI Token CounterCount tokens for GPT, Claude, Gemini, DeepSeek, Qwen and more.Tool
  • LLM API Cost CalculatorEstimate per-request, daily and monthly API costs.Tool
  • AI Model ComparisonCompare prices, context windows and features across models.Tool
  • AI Model Pricing PagesSpecs, real costs and cheaper alternatives for popular models.Tool
  • Context Window CheckerSee whether your text fits each model’s context window.Tool
  • Subscription vs API CalculatorFind out whether a chat plan or the API is cheaper for you.Tool
  • GPU / VRAM CalculatorCheck how much VRAM a local model needs and which GPUs fit.Tool
  • Prompt Caching CalculatorEstimate savings from prompt caching.Tool

MCP server · Reference servers

Fetch MCP server

Lets an assistant fetch a web page and read it as Markdown, in chunks, with private addresses blocked by default.

Maintainer
Model Context Protocol project (official)
Licence
MIT
Runs
Locally (uvx)
Sign-in
None
Last activity
2026-10-11

Overview

What it does

Fetch is a Python reference server from the MCP project. It has one tool, fetch, which downloads a URL and turns the HTML into Markdown, or returns the raw content if asked. Long pages are cut to a set length, and the model can ask for the next chunk with a start index, so it can read a page in parts until it finds what it needs. The same action is also offered as a prompt.

It suits quick lookups of a known URL when you don’t need a search engine or a full browser. By default it obeys robots.txt for model-initiated requests and refuses private, loopback, link-local and cloud metadata addresses, checking each redirect too. Flags let you ignore robots.txt, allow private addresses, set a custom user agent or use a proxy. Installing Node.js is optional and switches it to a different HTML simplifier.

Tools

What the assistant can do

  • Fetch a URL and convert the page to Markdown
  • Return raw content without conversion
  • Read long pages in chunks with max_length and start_index
  • Offer the same fetch as a user prompt

Install

Install Fetch MCP server in your app

You need:

  • uv (for uvx) or pip, with Python 3.10 or newer
  • Node.js (optional, for the alternative HTML simplifier)

Run in your project. --scope local keeps the servers private to you in this project; use user for every project.

Terminal
claude mcp add --scope local fetch -- uvx mcp-server-fetch
Native Windows (PowerShell)
claude mcp add --scope local fetch -- uvx mcp-server-fetch

Want it alongside other servers? Open Fetch MCP Server in the MCP config generator and add the rest; it writes one file per app. The snippets follow the setup in the server’s README, checked on 2026-10-14.

Before you install

  • It reaches any public host on the internet, so page content it returns can contain prompt-injection text.
  • Private, loopback and metadata addresses are refused by default; --allow-private-ips turns that guard off and the README notes it does not stop DNS rebinding.
  • --ignore-robots-txt makes it ignore site crawling rules; leave it off unless you have a reason.
  • A local MCP server runs with your user account’s permissions. Install it only from the source linked above. See how to choose a safe server.

Why it’s listed

Why we list it

The simplest way to let an assistant read a web page, with sensible network guards, kept by the MCP project.

Related

All MCP serversMCP config generatorGuide: MCP servers in Claude CodeGuide: what is MCP?