Stax
← All stacks

The Web Researcher Stack

Three servers an agent needs to actually read the web: a search engine, a scraper that handles JS, and a converter that turns whatever it pulls back into clean Markdown.

Use case: Build a research agent that searches, scrapes, and digests. Useful for competitive intel, lead research, or feeding fresh sources into a long writeup.

Install the whole stack

Pick the client you use. Each tab installs all 3 servers.

claude mcp add io-github-tavily-ai-tavily-mcp -- npx -y tavily-mcp
claude mcp add firecrawl-firecrawl-mcp-server -- npx -y firecrawl-mcp
claude mcp add microsoft-markitdown -- uvx markitdown-mcp

Servers in this stack

  • Tavily

    by tavily-ai · ★ 1,917

    Search built for LLMs, not humans. Returns clean results without ad cruft.

    claude mcp add io-github-tavily-ai-tavily-mcp -- npx -y tavily-mcp

  • Firecrawl

    by firecrawl · ★ 6,229

    Renders JS-heavy pages and returns clean Markdown. The thing you want when basic curl scraping returns an empty shell.

    claude mcp add firecrawl-firecrawl-mcp-server -- npx -y firecrawl-mcp

  • Markitdown

    by microsoft · ★ 121,110

    Converts PDFs, Word docs, slides, and images to Markdown. The bridge between 'I found a source' and 'I can quote it'.

    claude mcp add microsoft-markitdown -- uvx markitdown-mcp