Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
作者提供的说明(英文原文)
AI-agent website mirroring and offline archive browsing over MCP, powered by HTTrack.
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"httrack": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/techventures-studio/httrack-mcp-server:v1.0.3"
]
}
}
}在终端中运行这条命令
claude mcp add httrack -- docker run -i --rm ghcr.io/techventures-studio/httrack-mcp-server:v1.0.3粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"httrack": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/techventures-studio/httrack-mcp-server:v1.0.3"
]
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"httrack": {
"type": "stdio",
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"ghcr.io/techventures-studio/httrack-mcp-server:v1.0.3"
]
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
这个服务器在你的电脑上运行,而 ChatGPT 只能连接互联网上的服务器。请选择其他应用。
你的选择通过 Cookie 保存在这台设备上,无需账户。
这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
它在你的电脑上运行;安装后,工具列表会显示在你的 AI 应用中。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月6日
注册名称已通过 GitHub 账户 TechVentures-Studio 验证,该账户拥有此仓库
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
它在你的电脑上运行,没有互联网端点
在你的电脑上运行,改用本地设置
它在你的电脑上运行;我们只能读取互联网端点的工具列表
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。
它在你的电脑上运行。你的 AI 应用会在需要时启动它,数据不会经过作者的服务器(除非它自己调用在线服务)。
GitHub 仓库最后更新于–,最新版本(1.0.3)于35天前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。你的电脑上还需要有 Docker。
同一分类中有:Spec Driven Development, Remote Desktop Commander, Apify MCP Server
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
Extract data from any website with thousands of scrapers, crawlers, and automations on Apify Store ⚡
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
A FastMCP (Model Context Protocol) server that wraps the classic HTTrack website copier, so AI agents can mirror websites into a shared archive and then browse, read, and search the downloaded content — all over MCP.
Copyright (C) 2026 [Tech Ventures VCC](https://github.com/TechVentures-Studio). Licensed under the GNU Affero General Public License v3.0 (see LICENSE).
HTTrack is battle-tested for offline website copying, but it has no scripting API — only a GUI and a CLI. This server makes it agent-native: an LLM agent can call mirror_site to start a download in the background, poll mirror_status, and later answer questions from the archive via catalog, list_files, read_file, and search_files.
| Tool | Purpose |
|---|---|
mirror_site(url, project, depth, max_rate, filters) | Start a background mirror into the archive (/data/ ). One active mirror per project — a concurrent request for the same project returns already_running with the live job_id; different projects run in parallel. |
mirror_status(job_id) | Job status (running / done / failed + log tail). Omit to list recent jobs; also accepts the project name. |
catalog() | All projects: short name, short description (auto-derived from each site's /meta description), source URL, status, page/file counts, size. |
list_files(project, subpath) | Directory listing inside a project (name, type, size, modified). |
read_file(project, path, max_bytes) | Read one text file (HTML/JS/CSS/txt/json/md). Refuses binaries; caps content at 256 KB to keep agent context sane. |
search_files(project, query, max_results) | Case-insensitive substring search across a project's text files, capped results. |
Prerequisites: Docker.
docker build -t httrack-mcp-server .
docker run -d --name httrack-mcp --restart unless-stopped \ -p 9050:8000 \ -v /path/to/your/archive:/data:rw \ httrack-mcp
The container exposes the MCP server over streamable HTTP at http:// :9050/mcp.
Example for MCPHub-style gateways (streamable-http):
{
"httrack": { "type": "streamable-http", "url": "http://192.168.1.35:9050/mcp", "enabled": true, "description": "HTTrack site mirroring + archive browsing", "owner": "admin" } }
Clients that speak streamable HTTP directly can connect to /mcp with no extra auth (nobody is authenticated; run this on a trusted network, or put it behind your gateway's auth). See HELP.md for full usage, examples, and troubleshooting.
httrack process per project, isolated by its own output directory — no shared state between projects, so multiple agents can mirror different sites in parallel.mirror_site guards against double-mirroring the same project (in-process job table + a /proc scan for httrack processes writing to that project directory).GitHub Actions runs on every PR and push: Python syntax check of server.py, Dockerfile sanity assertions, and a Docker image build.
This program is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
If you run a modified version of this server as a network service, AGPL §13 requires you to offer your users the corresponding source of that modified version.
Copyright (C) 2026 Tech Ventures VCC. HTTrack itself is GPL-licensed by Xavier Roche and contributors; this project only shells out to it and is not affiliated with it.