What it does
Web Intel provides tools for reading and analysing web pages. read_url returns the main content as Markdown with navigation, footers and cookie banners removed. page_metadata returns the title, description, canonical URL, language, robots directives, Open Graph data, JSON-LD types, H1-H3 headings and image alt-text statistics. extract_links lists deduplicated internal and external links with their anchor text. tech_stack identifies frameworks, CMS, analytics, payments, CDN or hosting and security headers. seo_audit gives a 0-100 score, an A-F grade, 16 weighted checks and the top five fixes, and compare_seo ranks two to five pages against each other. Each paid tool costs 1 to 3 credits, while pricing and credits_balance are free.
It suits agents or developers who need structured page data or an SEO review of their own site compared with competitors. Failed fetches are refunded automatically, and invalid input is rejected before any credits are charged. Prices can be overridden through an environment variable.
Running it requires a KeyFlint license key. Setup differs by client: Claude Desktop accepts a one-click bundle, while Claude Code, Cursor and Windsurf use configuration entries. A hosted mode lets sellers run one server that bills many buyers over Streamable HTTP. In Claude, ChatGPT, VS Code and Goose, paid tools display result cards with credit balance and summary figures.
Limitations and requirements
Requires a KeyFlint license key; without a valid license, calls are rejected. Production use needs a KeyFlint license unless the use is non-commercial, under the Business Source License 1.1. Requests to private, loopback and link-local addresses are refused, and the check runs on every redirect. Responses are capped at 3 MB, requests time out after 15 seconds, and a maximum of 5 redirects is followed. Interactive result cards appear only in Claude, ChatGPT, VS Code and Goose; other clients receive plain JSON.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 15 Nov 2026 · its page on the source
Web Intel
Web pages to clean Markdown, metadata, links, tech stack and scored SEO audits. Pay per call.
v0.2.1As shown on registry.modelcontextprotocol.io when we last read it