rapid7-docs-mcp
A self-hostable MCP server that enables searching and reading Rapid7 documentation, blog posts, and resources through six tools, with optional authentication and Docker support.
README
Rapid7 Docs MCP Server
Search Rapid7 documentation, product info, blog posts, and resources from any MCP-compatible AI client. Self-hostable, authenticated, runs in Docker.
Built on FastMCP. Serves both Streamable HTTP and SSE — works with Claude Desktop, OpenChamber, Bifrost, and any MCP-compatible client.
Disclaimer: Vibe coded with Claude Code and Opencode. Created in personal time and is not officially supported or associated with Rapid7 and only uses public resources. Use at your own risk. Do not approach Rapid7 for support or issues regarding this project. Please open an issue instead.
Quick Start
Option 1: GHCR pre-built images
curl -O https://raw.githubusercontent.com/drunkrhin0/rapid7-docs-mcp/main/docker-compose.yml
echo 'IMAGE_REGISTRY=ghcr.io/drunkrhin0/' > .env
docker compose pull
docker compose up -d
Option 2: Build from source
git clone https://github.com/drunkrhin0/rapid7-docs-mcp
cd rapid7-docs-mcp
docker compose up -d
On first boot, the crawler indexes ~2,000 docs pages (10–30 min). After that, starts instantly. Data persists in Docker volumes. Cron keeps it fresh.
Connect any MCP client to: http://localhost:8002/mcp (Streamable HTTP)
SSE for OpenChamber/Bifrost: docker compose --profile sse up, then http://localhost:8004/mcp
How it works
Two services by default. SSE is opt-in (--profile sse):
| Container | What it does |
|---|---|
| crawler (Node.js) | Scrapes docs.rapid7.com, documentation.rapid7.com, extensions.rapid7.com, rapid7.com → markdown + JSON indexes |
| mcp-server (Python/FastMCP) | Serves 6 search tools via Streamable HTTP with optional API key auth |
| mcp-server-sse (opt-in) | SSE endpoint for clients that don't support Streamable HTTP |
Tools
| Tool | Description |
|---|---|
docs_search |
Full-text search with ranked results and snippets |
docs_read |
Read a page by path or URL |
docs_list |
Browse sections and page counts |
get_product_knowledge |
Product marketing, features, pricing, FAQs |
search_blog |
Search 3,600+ blog posts by keyword and category |
search_resources |
Search whitepapers, reports, guides |
Authentication (optional)
Set MCP_API_KEYS in .env to require API keys. Omit for open access.
MCP_API_KEYS=key1,key2,key3
OAuth providers (Auth0, Google, GitHub, etc.) supported via FastMCP auth docs.
Configuration
| Variable | Default | Description |
|---|---|---|
MCP_PORT |
8002 |
Streamable HTTP port |
SSE_PORT |
8004 |
SSE port (requires --profile sse) |
HEALTH_PORT |
8001 |
Health endpoint port |
MCP_API_KEYS |
(open) | Comma-separated API keys |
MCP_RATE_LIMIT |
60 |
Requests/min per key |
IMAGE_REGISTRY |
(empty) | GHCR prefix, e.g. ghcr.io/user/ |
GIT_SOURCE |
GitHub URL | Git repo for docker compose build |
GITHUB_TOKEN |
(empty) | GitHub PAT for higher API rate limits |
CRAWL_SECTIONS |
(all) | Space-separated section list |
CRAWL_SCHEDULE |
0 2 * * * |
Docs crawl cron |
CRAWL_EXTENSIONS |
true |
Crawl extensions site |
CRAWL_SITE |
true |
Crawl products, blog, resources |
CRAWL_EXTERNAL |
false |
Crawl GitHub + OpenAPI specs |
TZ |
UTC |
Cron timezone |
Full env var reference: docker-compose.yml.
Manual crawls
docker compose run --rm crawler npm run crawl # all docs
docker compose run --rm crawler npm run crawl -- --section insightidr # one section
docker compose run --rm crawler npm run crawl:extensions # extensions
docker compose run --rm crawler npm run crawl:site # products, blog, resources
docker compose run --rm crawler npm run crawl:external -- --insightvm-api # OpenAPI specs
Run npm run crawl -- --list inside the container to see available sections.
<details> <summary>Full crawl CLI reference</summary>
Docs: crawl -- --section insightidr | --url <url> | --list | --verbose
Site: crawl:site -- --products | --blog | --resources | --product command
External: crawl:external -- --metasploit | --velociraptor | --insightvm-api | --insightvm-cloud-api | --insightappsec-api | --insightidr-api | --insight-account-api | --credential-api | --insightconnect-api | --insightidr-detection-api
GitHub sources (Metasploit wiki, Velociraptor docs) fetched as markdown. OpenAPI specs split into one file per tag. Set GITHUB_TOKEN for higher API rate limits.
Cannot crawl: Threat Command / DRP API (requires authentication).
</details>
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。