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>