How to Research a Competitor Website with Claude and MCP
Use Claude and crawler-mcp to map a competitor site, fetch key pages, and summarise their strategy.
Competitor research usually means opening dozens of tabs and taking notes. With crawler-mcp, you can have Claude do the legwork: discover the site structure, fetch the pages that matter, and summarise findings in one conversation.
This guide shows a three-phase workflow using discover_links and fetch_page.
Step 1: Install crawler-mcp
Run the install script:
curl -fsSL https://install.crawler.sh/install-mcp.sh | shThis downloads the correct binary for your platform to ~/.crawler/bin/crawler-mcp.
For more detail, see the installation guide.
Step 2: Wire it into Claude Desktop
Follow the Claude Desktop setup guide to add crawler-sh to your config file and restart the app.
Step 3: Discover the site structure
Start by mapping the competitor’s site:
Use crawler-sh to discover_links on https://competitor.com to depth 2.
Claude returns a list of URLs with titles. Ask Claude to organise them by category:
Group those links by section - product pages, blog posts, pricing, support, etc.
Step 4: Fetch key pages
Pick the pages that matter most and fetch them:
Use crawler-sh to fetch https://competitor.com/pricing and https://competitor.com/features, then compare their positioning to ours.
Claude calls fetch_page for each URL and produces a side-by-side comparison.
Step 5: Summarise findings
After fetching the key pages, ask Claude to synthesise:
Based on the pages we fetched, summarise their messaging strategy, pricing tiers, and key differentiators.
Claude reasons over the fetched Markdown and gives you a structured competitor analysis.