Power your agent
Install from the publisher. Not a PlugMCP endpoint.
Host a copy with our keys and logs.
No packaged release in the registry yet. Follow the repository README.
Remote URL or a command on your machine? How the two installs differ
Run on PlugMCP
A new server you create. It is not this listing's official endpoint. Use the publisher install when you want their URL and their credentials.
Launchnxm-webtomd turns any web article into clean Markdown — right inside your AI agent.
Point it at a URL and it fetches the page, strips the navigation, ads, and boilerplate, and hands back the main article content as tidy Markdown (with optional YAML front‑matter). It speaks the Model Context Protocol (MCP) over stdio, so it plugs straight into agents like OpenCode, Claude Code, Cursor, and Kiro as a callable tool.
It is small, fast, and dependency‑light — a single self‑contained binary, no runtime downloads, nothing to configure beyond one line in your agent.
Fully open source (MIT / Apache‑2.0). Written in Rust.
From a noisy web page to clean, structured Markdown:

A real conversion, live in the terminal:

Agents are great at reasoning over text, but terrible at reading the raw web: a modern article page is 90% chrome — menus, cookie banners, share widgets, tracking scripts. Feeding all of that to a model wastes tokens and invites prompt‑injection from hidden page content.
nxm-webtomd does the dirty work first:
 references only — no image bytes are downloaded.Everything runs locally. Only the article you ask for is fetched; nothing else leaves your machine.
One command. It auto‑detects your system (macOS Apple Silicon or Linux x86_64), downloads the matching binary from the latest GitHub release, verifies its SHA‑256, and installs it to ~/.local/bin:
curl -fsSL https://raw.githubusercontent.com/dangranaz/nxm-webtomd/main/install.sh | sh
If ~/.local/bin is not on your PATH, add it:
export PATH="$HOME/.local/bin:$PATH"
Supported platforms: macOS arm64 (Apple Silicon) and Linux x86_64.
Requires a Rust toolchain (1.75+):
git clone https://github.com/dangranaz/nxm-webtomd.git
cd nxm-webtomd
cargo build --release
# binary at target/release/nxm-webtomd
nxm-webtomd is a standard MCP server over stdio, so you configure it like any other MCP tool — the agent manages the process lifecycle.
Here is an example for OpenCode — add it to your opencode.json (global) or opencode.jsonc under the mcp key:
{
"$schema": "https://opencode.ai/config.json",
"mcp": {
"nxm-webtomd": {
"type": "local",
"command": ["nxm-webtomd"],
"enabled": true
}
}
}
The configuration follows the same pattern in other agents (Claude Code, Cursor, Kiro…): a local MCP server whose command is nxm-webtomd over stdio. For example, in a mcpServers‑style config:
{
"mcpServers": {
"nxm-webtomd": {
"command": "nxm-webtomd",
"args": [],
"env": {}
}
}
}
Then just ask, in a prompt:
"use nxm-webtomd to fetch
https://example.com/articleand save it as Markdown to ./article.md"
The server exposes a single, focused MCP tool:
output_path → the Markdown comes back inline in the tool response.output_path → the Markdown is written to that file and the tool reports how many bytes it wrote.Fetching a dense technical article and converting it produces Markdown like this (front‑matter + body):
---
title: "Hexo Labs Open-Sources SIA: A Self-Improving Agent That Updates Both the Harness and the Model Weights"
canonical_url: "https://www.marktechpost.com/2026/05/29/hexo-labs-open-sources-sia-.../"
byline: "Asif Razzaq"
published_time: "2026-05-29T07:28:37+00:00"
---
Most AI agents stop improving once a human stops tuning them. The model is fixed.
The scaffold around it is fixed. Hexo Labs wants to move both at once...
## What is SIA (Self-Improving AI)
SIA splits a task-specific agent into two parts. The first is the harness...
| Task | Initial | Prev. SOTA | SIA-H | SIA-W+H |
| ---- | ------- | ---------- | ----- | ------- |
| LawBench (top-1 acc) | 13.5% | 45.0% | 50.0% | 70.1% |
Tables, headings, links, and code blocks survive the trip; navigation, ads, and scripts do not.
The server reads one JSON‑RPC object per line on stdin and replies on stdout:
printf '%s\n%s\n' \
'{"jsonrpc":"2.0","id":1,"method":"initialize","params":{}}' \
'{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}' \
| nxm-webtomd
Call the tool directly:
{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{
"name":"web_to_md",
"arguments":{"url":"https://example.com/article","output_path":"./article.md"}}}
If nxm-webtomd saves you time or tokens, please give the repository a star and share it — it is the simplest way to help the project reach other developers. Feedback and suggestions are welcome via issues.
Licensed under either of
at your option.
A small, dependency-light MCP server that fetches web articles and converts them to clean Markdown (stdio).
Tools
No tool list yet. Remote servers get one after the first health check.
FAQ
Copy the stdio config for Cursor, Claude Desktop, Claude Code, or VS Code on this page. It uses the publisher's install, not a PlugMCP proxy.
No required environment variable is listed on this record.
Nxm Webtomd runs locally over stdio.
Last checked 2026-10-28. A small, dependency-light MCP server that fetches web articles and converts them to clean Markdown (stdio).