I Built an Automated Pipeline That Discovers New MCP Tools Every 6 Hours

The Problem with MCP Discovery

Model Context Protocol (MCP) is exploding. New servers appear daily on GitHub. But nobody has time to scroll through hundreds of repos looking for quality tools.

So I automated it.

The Stack

• 7 Python agents for collection, analysis, and content generation
• Next.js 16 frontend with Tailwind v4
• Vercel for deployment
• GitHub API for tool discovery

How It Works

1. Collector Agent crawls GitHub for repos matching mcp-server or model-context-protocol
2. Analyzer Agent scores each tool on quality, docs, activity
3. Writer Agent generates deep review articles for top tools
4. Publisher Agent pushes updates to the site
5. Ops Brain coordinates everything, generates strategy reports

Results

• 900+ tools cataloged
• 19 categories
• Updated every 6 hours automatically
• Zero manual work after setup

Try It

https://mcphello.com — free, no signup.

---

Full automation with Python agents. Deployed on Vercel.