Huoshui Fetch

io.github.huoshuiai42Published by GitHub account huoshuiai42v1.0.0Updated Oct 24, 2026

An MCP server that provides tools for fetching, converting, and extracting data from web pages.

VerifiedSTDIODesktop onlyWeb Search & ScrapingFiles & Storage

Overview

AI-generated overview

A local MCP server that lets an assistant fetch web pages and convert or extract their content into Markdown, text, links, metadata and images.

What you can do
Fetch pages from URLs, with configurable timeout, redirect handling, user agent and custom request headers for authenticated requests. Convert HTML or JSON into clean Markdown, plain text or sanitized HTML. Extract the main article body, links with filtering, page metadata such as title, description and OG tags, images with size filtering, and structured data from JSON-LD and microdata.
When to use it
Useful when an assistant needs to read web pages, turn HTML into readable Markdown or text, pull article content out of a page, or collect links, metadata, images and structured data for research or content work.
Requirements
Runs locally over stdio as a Python package (huoshui-fetch), requiring Python 3.11 or newer and its dependencies; typically launched with uvx or python -m. Network access is needed to fetch remote URLs. No accounts, API keys or environment variables are declared.
Before you install
fetch_with_headers accepts custom request headers, so credentials or tokens passed that way go to the target site; avoid sending secrets to untrusted URLs. Fetched page content is untrusted input and may contain misleading text. The server only reads pages; it does not write or change remote data.

Installation

In SourceWeft

  1. Open Huoshui Fetch in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

huoshui-fetch

A dedicated web content fetching and conversion MCP (Model Context Protocol) server that provides tools for fetching, converting, and extracting data from web pages.

Features

Fetching Tools

  • fetch_url: Fetch content from URLs with customizable timeout, redirect handling, and user-agent
  • fetch_with_headers: Fetch URLs with custom headers for authenticated requests

Conversion Tools

  • html_to_markdown_tool: Convert HTML to clean Markdown format
  • html_to_text_tool: Extract plain text from HTML
  • clean_html_tool: Remove scripts/styles and sanitize HTML
  • json_to_markdown_tool: Convert JSON data to readable Markdown

Extraction Tools

  • extract_article_tool: Extract main article content using readability
  • extract_links_tool: Extract all links with filtering options
  • extract_metadata_tool: Extract page metadata (title, description, OG tags)
  • extract_images_tool: Extract images with size filtering
  • extract_structured_data_tool: Extract JSON-LD and microdata

Installation

From MCP Registry (Recommended)

This server is available in the Model Context Protocol Registry. Install it using your MCP client.

mcp-name: io.github.huoshuiai42/huoshui-fetch

bash
# Using uv (recommended)uv sync
# Or install from GitHubpip install git+https://github.com/yourusername/huoshui-fetch.git

Usage

Run with uvx (recommended for one-time use)

bash
# From the repositoryuvx --from . huoshui-fetch
# From GitHub (once published)uvx --from git+https://github.com/yourusername/huoshui-fetch.git huoshui-fetch

Run directly

bash
# Using uvuv run python -m huoshui_fetch
# Or if installedpython -m huoshui_fetch

The server communicates via standard input/output, making it perfect for integration with Claude Desktop and other MCP-compatible clients.

Configuration for Claude Desktop

Add to your Claude Desktop configuration:

json
{  "mcpServers": {    "huoshui-fetch": {      "command": "uvx",      "args": ["--no-cache", "--from", ".", "huoshui-fetch"],      "cwd": "/path/to/huoshui-fetch"    }  }}

Or if installed from GitHub:

json
{  "mcpServers": {    "huoshui-fetch": {      "command": "uvx",      "args": [        "--from",        "git+https://github.com/yourusername/huoshui-fetch.git",        "huoshui-fetch"      ]    }  }}

Example Usage

Once configured, you can use the tools in Claude Desktop:

// Fetch a webpagefetch_url("https://example.com")
// Convert HTML to Markdownhtml_to_markdown_tool("<h1>Hello</h1><p>World</p>")
// Extract article contentextract_article_tool(html_content, "https://example.com/article")

Requirements

  • Python 3.11+
  • Dependencies listed in pyproject.toml

Development & Publishing

This project includes comprehensive automation for building and publishing to PyPI.

Automated Publishing Workflow

bash
# Complete automated workflow (TestPyPI + PyPI)uv run python scripts/publish.py --include-pypi
# TestPyPI only (recommended for testing)uv run python scripts/publish.py
# Bump version and publishuv run python scripts/publish.py --version-bump patch --include-pypi

Individual Commands

bash
# Version managementuv run python scripts/version_manager.py --checkuv run python scripts/version_manager.py --bump patch
# Setup PyPI credentials (first time)uv run python scripts/credentials_setup.py
# Build packageuv run python scripts/build.py
# Run comprehensive testsuv run python scripts/test.py
# Upload to PyPIuv run python scripts/upload.py

Features

  • ✅ Version Management: Automatic synchronization across all files
  • ✅ Quality Checks: Ruff linting and MyPy type checking
  • ✅ Build Automation: Clean builds with validation
  • ✅ Testing Suite: Comprehensive package and functionality tests
  • ✅ Publishing Workflow: TestPyPI → PyPI using uv publish (supports .pypirc files)
  • ✅ Error Recovery: Built-in error handling and recovery options

See PUBLISHING.md for detailed documentation.

DXT Extension

This project supports DXT (Desktop Extensions) format for easy distribution and installation.

To build the DXT extension:

bash
python build_dxt.py

This will create a huoshui-fetch-{version}.dxt file that can be installed in compatible AI desktop applications.

License

MIT

Source: README.md at commit 375d268

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v1.0.0LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts