World Monitor
Real-time global intelligence on conflict, markets, energy, trade, shipping and cyber. 104 tools.
Route long-tail MCP servers through three lazy tools and reviewed declarative Methods.
选择你使用的 AI 应用,然后复制设置。
通过 3/3 项检查 · 在你的电脑上运行
第一步:你用哪个 AI?
选择你用的应用,步骤会显示在下方。
可连接 Claude,但需要先准备
需要先在电脑上安装 Node.js(只需一次) · 仅限电脑上的 Claude 应用(Windows 或 Mac),网页和手机不可用
SettingsDeveloperEdit Config
{
"mcpServers": {
"metamcp": {
"command": "npx",
"args": [
"-y",
"@mentu/metamcp"
]
}
}
}用 MetaMCP 告诉我你能帮我做什么,然后做一件给我看✓ 如果 Claude 说正在使用 MetaMCP,或回答中出现 MetaMCP 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 Anthropic 帮助页面核实 · 你的界面可能因版本略有不同
ChatGPT 暂时无法连接 MetaMCP
MetaMCP 需要安装在你的电脑上,而 ChatGPT 只能连接在网上运行的服务器
A · 使用能连接 MetaMCP 的应用
B · 其他 类中能连接 ChatGPT 的其他服务器
从同类中使用较多的里挑选,仅限用地址即可连接、无需密钥的
可连接 Claude Code,但需要先准备
需要先在电脑上安装 Node.js(只需一次)
Mac:按 Command+空格,输入 Terminal · Windows:打开 PowerShell
claude mcp add --scope user metamcp -- npx -y @mentu/metamcp应能看到 MetaMCP 在列表中并显示已连接
用 MetaMCP 告诉我你能帮我做什么,然后做一件给我看✓ 如果 Claude Code 说正在使用 MetaMCP,或回答中出现 MetaMCP 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 Anthropic 帮助页面核实 · 你的界面可能因版本略有不同
可连接 Cursor,但需要先准备
需要先在电脑上安装 Node.js(只需一次)
用 MetaMCP 告诉我你能帮我做什么,然后做一件给我看✓ 如果 Cursor 说正在使用 MetaMCP,或回答中出现 MetaMCP 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 Cursor 帮助页面核实 · 你的界面可能因版本略有不同
可连接 VS Code,但需要先准备
需要先在电脑上安装 Node.js(只需一次)
用 MetaMCP 告诉我你能帮我做什么,然后做一件给我看✓ 如果 VS Code 说正在使用 MetaMCP,或回答中出现 MetaMCP 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 VS Code 帮助页面核实 · 你的界面可能因版本略有不同
Claude Code
claude mcp add --scope user metamcp -- npx -y @mentu/metamcpClaude Desktop (claude_desktop_config.json)
{
"mcpServers": {
"metamcp": {
"command": "npx",
"args": [
"-y",
"@mentu/metamcp"
]
}
}
}Cursor (~/.cursor/mcp.json)
{
"mcpServers": {
"metamcp": {
"command": "npx",
"args": [
"-y",
"@mentu/metamcp"
]
}
}
}VS Code (mcp.json)
{
"servers": {
"metamcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@mentu/metamcp"
]
}
}
}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。
依据这个服务器的连接方式和各应用自己的文档,检查于 2026年11月10日。
在你的电脑上运行,适用于
ClaudeGooseLM StudioMsty StudioCherry StudioCursorVS Code + GitHub CopilotVisual StudioWindsurf (Devin Desktop)ZedKiroClineContinueJetBrains AI AssistantClaude CodeCodex CLIGemini CLIOpenCodenodecodeWarpLibreChatRaycast
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12日
注册名称已通过 GitHub 账户 dataware-co 验证,该账户拥有此仓库
代码最后更新于43天前
使用 Apache-2.0
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
提问、分享技巧或报告无法使用。按“有帮助”数量排序。
还没有评价
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 3 项适用的基础检查中通过了 3 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43天前,最新版本(1.0.0)于3个月前进入注册中心。
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 Node.js。
同一分类中有:World Monitor, Claude Flow, Windmill
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:↓ 每周下载量i★ 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
Real-time global intelligence on conflict, markets, energy, trade, shipping and cyber. 104 tools.
AI orchestration with hive-mind swarms, neural networks, and 87 MCP tools for enterprise dev.
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。
MetaMCP is a secure, on-demand gateway for the long tail of MCP servers. It gives an MCP client three stable tools:
mcp_discover finds configured servers, cached tool schemas, and reviewed Methods without starting every child.mcp_call lazily calls one explicitly named child tool.mcp_run executes a bounded, schema-validated declarative Method.MetaMCP is deliberately not a replacement for every direct MCP connection. Keep important, frequently used, compact, or strongly authenticated MCPs direct. Put irregular long-tail servers behind MetaMCP, and promote repeated multi-step rituals into Methods.
┌─ direct: GitHub / Codex Apps / core runtime
MCP client ────────────────────┤ └─ MetaMCP (3 tools) ├─ discover cached capabilities ├─ call one lazy child └─ run reviewed Methods
| Path | Best fit | Why |
|---|---|---|
| Direct MCP | High-frequency, compact, security-sensitive, or foundational servers | Preserves typed schemas, native auth, and explicit approvals |
mcp_discover + mcp_call | Long-tail or irregular capabilities | Keeps the client surface small without hiding the assembly language |
mcp_run | Repeated Acquire → Normalize → Analyze workflows | Makes bounded behavior testable, versioned, and evidence-producing |
Do not route billing, infrastructure mutation, identity, or another high-consequence server through MetaMCP merely to reduce tool count. The right boundary is operational, not ideological.
Requires Node.js 20 or newer.
npx @mentu/metamcp@latest --config .mcp.json
Inspect the complete model-facing surface before configuring a client:
npx @mentu/metamcp@latest tools
npx @mentu/metamcp@latest tools --json
The inspector reads the same definitions returned by MCP tools/list, then exits before loading configuration, opening storage, starting a child, or binding a transport. --json includes the complete input schemas for automated review and version-to-version diffs.
Create .mcp.json:
{
"mcpServers": { "filesystem": { "command": "npx", "args": ["-y", "@modelcontextprotocol/server-filesystem@2026.7.10", "/path/to/allowed/files"] }, "internal-api": { "command": "node", "args": ["./servers/internal-api.js"], "env": { "API_TOKEN": "${INTERNAL_API_TOKEN}" }, "inheritEnv": ["HTTP_PROXY"] } } }
Child servers start only when explicitly refreshed, called, or used by a Method. Plain discovery reads configuration and cached schemas; it does not spawn all children.
Server names are stable cache identities and must contain 1-128 letters, numbers, dots, underscores, or hyphens; path separators and traversal-like names are rejected.
init is preview-only unless --yes is supplied. Without a named client it considers existing client config files only.
metamcp init # preview, no writes
metamcp init --client Codex # preview one client metamcp init --client Codex --yes # apply atomically and write a .bak
Malformed JSON is rejected and left untouched. A named client may be created explicitly; MetaMCP never creates every supported client config by default.
For a manual client configuration, use an absolute path so the gateway does not depend on the client's working directory:
{
"mcpServers": { "metamcp": { "command": "npx", "args": [ "-y", "@mentu/metamcp@latest", "--config", "/absolute/path/to/.mcp.json" ] } } }
@latest is convenient for evaluation. Pin @mentu/metamcp@1.0.0 in controlled environments so upgrades are deliberate and reviewable.
{ "query": "capture screenshot", "kind": "tool" }
Discovery searches only live or cached schemas. To refresh one server from its live tool list:
{ "server": "browser", "refresh": true }
refresh without a server is rejected so an agent cannot accidentally fan out across the whole configuration.
{
"server": "browser", "tool": "capture_page", "args": { "url": "https://example.com" }, "timeoutMs": 60000 }
MetaMCP never automatically replays a child call after a timeout or transport failure. The child may have completed a mutation before the response was lost. A later Method may retry only when its manifest explicitly declares that step idempotency: "safe".
Put JSON manifests in .metamcp/methods/ or pass --methods . The child server and tool names below are illustrative; bind them to reviewed servers in your own config:
{
"apiVersion": "metamcp.io/v1alpha1", "kind": "Method", "metadata": { "name": "content.acquire-and-normalize", "version": "1.0.0", "description": "Acquire content and normalize it into a stable record" }, "spec": { "effects": "read", "inputSchema": { "type": "object", "properties": { "url": { "type": "string" } }, "required": ["url"], "additionalProperties": false }, "steps": [ { "id": "acquire", "server": "fetch", "tool": "fetch", "args": { "url": "${input.url}" } }, { "id": "normalize", "server": "content", "tool": "normalize", "dependsOn": ["acquire"], "args": { "document": "${steps.acquire.structuredContent}" } } ], "output": "${steps.normalize.structuredContent}" } }
Then call:
{ "method": "content.acquire-and-normalize", "input": { "url": "https://example.com" } }
Methods are declarative rather than arbitrary JavaScript. They have bounded step counts, deadlines and output sizes; input/output JSON Schemas; explicit read/write effects; safe interpolation; typed gaps; and a per-step trace. Write or mixed-effect Methods are disabled unless the gateway operator starts MetaMCP with --allow-writes.
See Method Mode, the manifest schema, and the example Method. The design generalizes the consistency layer documented by Crawlio Method Mode.
${NAME} references in env and HTTP headers resolve from the host environment by default. An unresolved reference fails startup; it is never passed to a child as a literal placeholder.
MetaMCP does not copy its ambient environment into stdio children. It inherits only a small runtime allowlist (PATH, home/temp/locale variables, and platform equivalents), variables named in inheritEnv, and values explicitly set in the child env block. Embedders can install a custom SecretProvider for a keychain or vault.
Discovery is local keyword search by default. To opt into Voyage-backed semantic search, set METAMCP_VOYAGE_API_KEY explicitly; discovery queries will then be sent to Voyage and the optional local SQLite vector index will be enabled. Ambient ANTHROPIC_API_KEY or VOYAGE_API_KEY variables never activate network calls.
Remote child servers use url, transportType, headers, and the existing OAuth fields:
{
"mcpServers": { "remote": { "url": "https://mcp.example.com/mcp", "transportType": "http", "headers": { "Authorization": "Bearer ${REMOTE_TOKEN}" } } } }
HTTP mode binds to 127.0.0.1 by default:
metamcp --transport http --port 8080 --config .mcp.json
An unauthenticated non-loopback bind fails closed. Configure OAuth resource-server validation or METAMCP_HTTP_BEARER_TOKEN before exposing the listener. Browser requests with an Origin header are denied unless the exact origin is supplied with --allow-origin or METAMCP_ALLOWED_ORIGINS.
MetaMCP serves legacy MCP clients and the 2026-07-28 stateless request envelope over stdio and Streamable HTTP. See Architecture for the supported boundary and deployment guidance.
Completed mcp_call and mcp_run attempts are serialized into .metamcp/ledger.jsonl. Export a portable hash-linked bundle:
metamcp export-evidence \
--ledger .metamcp/ledger.jsonl \ --out .metamcp/evidence-bundle.json
metamcp export-evidence --out .metamcp/evidence-bundle.json --verify
The operational ledger is not a remote attestation system. The export detects later changes inside a bundle; it does not prove that a compromised host recorded every event.
The package still ships a human-operated server gallery:
metamcp add --list
metamcp add playwright sentry --config .mcp.json
The runtime never installs packages in response to an MCP tool call. Installation remains an explicit CLI/user action.
Gallery entries are either a local launcher (command + args) or a hosted Streamable HTTP endpoint (url, shown as 🌐 in --list). Hosted entries are written to .mcp.json as native remote servers; the gallery never routes a remote server through a stdio bridge package.
Version 1.0 intentionally removes the model-facing provisioning, skill-advice, and JavaScript execution tools. It also changes HTTP binding, child environment inheritance, retries, and init. Read Migration to 1.0 before upgrading.
Child MCP servers are trusted local or remote code with their own permissions. MetaMCP is a policy and lifecycle boundary, not an OS sandbox for untrusted packages. Review commands, pin packages where appropriate, scope credentials per child, and keep dangerous direct servers behind client-side human approval.
Report vulnerabilities privately as described in SECURITY.md.
npm ci
npm run typecheck npm test ./scripts/smoke-test.sh npm run check:release npm pack --dry-run
npm publish runs the full verify:release gate again. The gate derives the public tool surface from the built CLI and checks package, lockfile, changelog, and official MCP Registry metadata for version drift.
Apache-2.0 licensed. Maintained by Mentu AI.