ActableSite AI Crawler Monitor

com.actablesite由 actablesite.com 发布官方发布者v1.4.1更新于 Oct 24, 2026

Check AI crawler robots.txt policy and monitor public-site policy, sitemap, and llms.txt changes.

发布方自己的服务：ActableSite AI Crawler Monitor

已验证Streamable HTTP可网页运行网页搜索与抓取安全与监控商业与电商

服务详情

标识符: com.actablesite/readiness
传输: Streamable HTTP
运行时: Web executable
信任: Verified
发布于: Sep 16, 2026

安全提示

MCP 服务会接收对话中的工具调用请求，并可能执行外部操作。启用前请先了解该服务能做什么，并只授予它所需的权限。

注册中心已验证此服务的命名空间属于 actablesite.com，且该域名同时运行其端点。这可以确认发布者，但并不代表该服务的工具是安全的。

SourceWeft 的检查只涵盖条目是否完整、端点是否响应；不审查服务器的代码，也不审查其工具的行为。

在 README 中添加徽章

向读者展示该 MCP 服务已收录于 SourceWeft，徽章链接到本页。

[![Listed on SourceWeft](https://sourceweft.com/badge/mcp/com.actablesite%2Freadiness.svg)](https://sourceweft.com/mcp/com.actablesite%2Freadiness?utm_source=readme&utm_medium=badge)

概览

AI 生成的概览

一个远程只读 MCP 服务器，用于审计公开网站的 AI 爬虫策略与发现文件，并返回付费监控服务的报价信息。

可以做什么: 审计公开网站的身份信息、报价证据、操作路径、发现文件、元数据和结构化数据。评估 robots.txt 中针对八个 OpenAI、Anthropic、Perplexity 和 Google 爬虫令牌的规则。读取按月爬虫监控、一次性报告和每周实践雷达版的信息性报价详情，包括价格、范围、频率和限制。
适用场景: 当你希望助手检查公开网站的 robots.txt 和发现文件是否允许主流 AI 爬虫，或查看网站公开展示的内容时适用。也适合在决定之前了解服务商付费监控方案的具体条款。
运行要求: 远程 Streamable HTTP 端点，无需身份验证。可通过 Node.js 20 或更高版本运行本地 stdio 桥接，或使用公开容器镜像。仅接受公开的 HTTP 和 HTTPS 网站，私有、本地和保留地址会被拒绝。

安装前请注意

报价工具仅提供信息，无法打开结账或购买任何内容；任何购买都需要用户明确确认。结果不保证被抓取、收录、引用、排名、流量或法律合规，付费监控使用合成检查，不能证明真实爬虫活动。每个调用方每小时限 60 次调用。

安装

在 SourceWeft 中

打开控制台中的 ActableSite AI Crawler Monitor，将其添加到工作区。
为需要使用其工具的对话启用该服务。

Web executable，通过 Streamable HTTP。远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体，或 SourceWeft 对话中。智能体会读取该服务器的安装指南，先向你展示发布者、端点和登录方式，经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/com.actablesite%2Freadiness/install.md，并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前，先告诉我它的发布者、端点和登录方式，等我确认。不要把 API 密钥写进命令或我会提交的文件；问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。

{
  "mcpServers": {
    "actablesite-ai-crawler-monitor": {
      "type": "http",
      "url": "https://actablesite.com/mcp"
    }
  }
}

README

ActableSite MCP

A remote, read-only MCP server for public website evidence, AI crawler policy, and explicit Crawler Watch, report, and Practice Radar offer metadata.

Tools

It exposes five tools over Streamable HTTP:

audit_public_website checks public identity, offer evidence, action paths, discovery files, metadata, and structured data.
check_ai_crawler_policy evaluates eight OpenAI, Anthropic, Perplexity, and Google crawler tokens.
get_crawler_watch_offer returns the $9 monthly price, one-site scope, 15-minute cadence, confirmation rule, delivery, cancellation, and synthetic-check limitations. It cannot open checkout or purchase anything.
get_full_report_offer returns explicit metadata for an optional $19 one-time report. It cannot open checkout or purchase anything.
get_practice_radar_offer returns the current weekly-edition receipt, public sample, $39 monthly price, delivery, cancellation, and material NPI limitations. It cannot open checkout or purchase anything.

Connect

Claude Code:

For another MCP client, add a remote server named actablesite with:

URL: https://actablesite.com/mcp/
Transport: Streamable HTTP
Authentication: none

For clients that require a local stdio command, clone this repository and run the dependency-free bridge:

The bridge contains the visible MCP implementation source in src/stdio.mjs. Distribution version 1.4.2 exposes the same five read-only tools and sends requests only to ActableSite's bounded public JSON APIs. It has no file, shell, browser, account, or write capability.

The same stdio process is available as a public, non-root container:

The hosted server implements MCP protocol version 2025-06-18 and is published in the official registry as com.actablesite/ai-crawler-robots-monitor version 1.4.2.

Discovery metadata

Directory scanners can read either machine-readable discovery surface:

/.well-known/mcp.json identifies the official registry name, repository, remote endpoint, transport, authentication mode, and tool names.
/.well-known/mcp/server-card.json provides static tool descriptions and JSON input schemas for scanners that cannot complete a live MCP handshake.

Both files describe the same unauthenticated endpoint and five-tool inventory verified by this repository.

Ongoing crawler monitoring

The MCP tools report current public evidence and do not create background jobs. Crawler Watch is the separate $9/month service for one public website: it checks crawler policy, synthetic homepage responses, sitemap.xml, and llms.txt every 15 minutes and emails after a changed state appears twice.

Its synthetic requests do not authenticate provider IP ranges or prove crawling, indexing, citation, ranking, or traffic.

Safety contract

Public HTTP and HTTPS websites only; private, local, loopback, and reserved network destinations are rejected.
No browser control, file access, account access, write, or mutation tools.
No authentication header or API key is requested.
Tool calls are limited to 60 per caller per hour.
All three offer tools are informational and require explicit user confirmation before any separate checkout or purchase action.
Practice Radar records do not prove licensure, active operation, demand, or buying intent and must be verified before contact.
Results do not guarantee crawling, indexing, citation, ranking, recommendation, traffic, legal compliance, or revenue.

Verify the live server

Node.js 20 or newer is required.

The verifier initializes the live server, checks its protocol and version, and confirms the exact read-only tool inventory. It does not call an audit tool or create a funnel event.

Repository scope

This repository contains a runnable stdio implementation, public registry manifest, connection contract, live verifier, and security policy for the hosted ActableSite MCP server. The production runtime is operated at https://actablesite.com/mcp/.