Set it up
-
Turn on Fetch in your AI
In Claude: open Settings → Connectors and add Fetch. In ChatGPT: Settings → Apps & Connectors.
-
Add it to your app once
Copy the short config below into your app. It just tells your AI where to find Fetch.
-
Just ask
Tell it what you want in plain words — no special commands.
For example, say
“Fetch https://example.com and tell me what the page is about”
Need the exact config? Show it +
Add to your client
{
"mcpServers": {
"fetch": {
"command": "uvx",
"args": ["mcp-server-fetch"]
}
}
}Paste into your client's MCP configuration file.
Ready-to-run prompts
If you ask, Fetch does it. Copy one, paste it into Claude, Codex or ChatGPT, done. No setup.
Fetch Read and summarize a page
If you ask
“Fetch https://example.com/article and summarize the main points”
Then
Fetch a URL and summarize or extract content.
Fetch Inspect an API response
If you ask
“Fetch https://api.github.com/repos/anthropic/anthropic-sdk-python and show me the star count and language”
Then
Hit a public API endpoint and inspect the response.
Fetch Convert a page's content
If you ask
“Fetch this blog post and convert it to a clean markdown file”
Then
Grab a page and have AI convert it to another format.
Good to know
- · JavaScript-heavy pages return mostly empty content — fetch only sees initial HTML
- · Some sites block bot user agents — requests may fail or return error pages
- · Binary content (images, PDFs) returns raw bytes which may blow context windows
When to use
- ✓ You want to read a web page or API response in conversation
- ✓ The URL returns server-rendered HTML, JSON, or plain text
- ✓ You need a lightweight, free option for single-page fetches
- ✓ You're combining with other tools and just need the page content
When not to use
- × The page needs JavaScript to render content — use puppeteer or browserbase
- × You're crawling many pages — use firecrawl for bulk
- × You need authenticated fetches with complex cookie flows — use a proper client
About
Imagine telling your AI: 'Read the content of this URL and summarize it' — and it just fetches the page. This server gives your AI a simple tool to download a URL and return its content as text, like a basic browser that doesn't run JavaScript.
The official Fetch MCP server enables AI assistants to fetch and process web content. It handles HTML-to-Markdown conversion, follows redirects, and can extract specific content from pages.
A foundational building block for any AI workflow that needs to access web content, documentation, or APIs.
Workflows That Use Fetch MCP
Recommended recipe
Let Claude search and browse the live webReal-time search, page scraping, and full browser automation — Claude works with what's on the web right now.
Works With
Also Consider
FAQ
What does Fetch MCP do? +
Imagine telling your AI: 'Read the content of this URL and summarize it' — and it just fetches the page. This server gives your AI a simple tool to download a URL and return its content as text, like a basic browser that doesn't run JavaScript.
Do I need to know how to code? +
No. Turn it on in your AI's settings and ask in plain English — no terminal, no coding.
When should I use Fetch MCP? +
Reach for it when you need to: You want to read a web page or API response in conversation; The URL returns server-rendered HTML, JSON, or plain text; You need a lightweight, free option for single-page fetches; You're combining with other tools and just need the page content.
When should I avoid Fetch MCP? +
Skip it when: The page needs JavaScript to render content — use puppeteer or browserbase; You're crawling many pages — use firecrawl for bulk; You need authenticated fetches with complex cookie flows — use a proper client.
How do I install Fetch MCP? +
In Claude Code, run: claude mcp add fetch -- uvx mcp-server-fetch. Copy-paste configs for Claude Desktop, Cursor, and Windsurf are at the top of this page.
