MCPpedia

MCPpedia

The encyclopedia for MCP servers. Browse, compare, and evaluate every Model Context Protocol server we can find (36,000+ and growing — the live count is on the homepage) with transparent, automated scoring.

Live at mcppedia.org

Features

• Automated scoring — every server scored 0-100 across security, maintenance, efficiency, documentation, and compatibility, with color-coded grade badges (A/B/C/D/F)
• Real CVE scanning — queries OSV.dev for known vulnerabilities in npm and PyPI packages
• Token efficiency measurement — measures actual context window cost of each server's tool schemas
• Registry sync — pulls from the official MCP Registry daily
• Advanced filtering — filter servers by score tier, health status, transport type, and category; sort by score, stars, downloads, activity, or name
• One-click install configs — copy-paste configs for Claude Desktop, Cursor, Claude Code, and Windsurf with a copy button on every card
• My Servers — save favorites with heart icons, manage them on a personal dashboard, and export all configs as one JSON
• Trending — "Trending this week" widget on the homepage highlighting top-downloaded high-scoring servers
• Server comparison — side-by-side comparison of any two servers
• Server README rendering — rendered markdown README directly on server detail pages
• Popular searches — quick-access search suggestions shown on focus
• Community reviews and discussion — user reviews with star ratings, plus "works for me" verifications from real users
• Karma and contributor profiles — earn points for submitting servers, proposing edits, posting in discussions, and verifying servers; progress tiers on public /profile/[username] pages
• Sign in with GitHub or Google — OAuth via Supabase with a one-time /welcome step to pick a username
• Blog — auto-generated and editorial articles about MCP servers and the ecosystem, with sticky table of contents on wide screens
• Newsletter — weekly digest of new and trending servers
• Best-of lists — curated lists of top servers by use case (developers, AI agents, databases, security, monitoring, and more)
• Skills directory — curated catalog of Claude Code skills and plugins at /skills, browsable by category and compatible agent
• Embeddable badges — SVG score badges via /api/widget/[slug] in flat and detailed styles
• Public API — GET /api/v1/servers with full filtering, sorting, pagination, and CORS headers
• MCP server — an MCP server for querying MCPpedia data programmatically (mcppedia-server/)
• 3-step submit flow — guided server submission with repository, details, and classification steps

Tech Stack

• Framework: Next.js (App Router)
• Styling: Tailwind CSS
• Database: Supabase (PostgreSQL)
• Security data: OSV.dev API
• Server registry: registry.modelcontextprotocol.io
• AI: Anthropic Claude (schema extraction, blog generation, description enrichment)
• Hosting: Cloudflare Workers (via @opennextjs/cloudflare) — see docs/CLOUDFLARE.md

Getting Started

Prerequisites

• Node.js 20+
• A Supabase project
• GitHub personal access token (for bot metadata fetching)

Setup

git clone https://github.com/BbekShr/MCPpedia.git
cd MCPpedia
npm install
cp .env.example .env.local
# Fill in your keys in .env.local

Database

npx supabase link --project-ref your-project-ref
npx supabase db push

Run

npm run dev

Open http://localhost:3000.

Project Structure

app/              Next.js pages and API routes
  api/v1/         Public REST API
  api/widget/     Embeddable SVG badge endpoint
bots/             Automated data pipeline scripts (run via GitHub Actions)
components/       React components
lib/              Shared utilities, types, scoring engine
scripts/          One-off data scripts
supabase/         Database migrations
content/          Static content (blog posts, guides)
data/             Generated data (comparison pairs, etc.)
cli/              CLI tool for querying MCPpedia
mcppedia-server/  MCP server for querying MCPpedia data

Bots

Automated bots run via GitHub Actions to keep data fresh:

Bot	Schedule	What it does
sync-registry	Daily 4am UTC	Syncs servers from the official MCP Registry
discover	Daily 5am UTC	Discovers new MCP servers from GitHub and npm
update-metadata	Daily 6am UTC	Refreshes GitHub stars, commits, issues, npm downloads
extract-install-info	Daily 7am UTC	Extracts install configs and categories
enrich-descriptions	Daily 7:30am UTC	Enriches server descriptions using Claude
compute-scores	Daily 8am UTC	Computes all scores (security, efficiency, docs, etc.)
extract-schemas	Daily 8am UTC	Extracts tool schemas from server packages using Claude
snapshot-metrics	Daily 8:30am UTC	Snapshots ecosystem metrics for historical tracking
track-trending	Daily 9:30am UTC	Tracks trending servers
blog-security-alerts	Daily 9:30am UTC	Generates blog posts for security alerts
publish-scheduled-blog	Daily 12pm UTC	Publishes scheduled blog posts
generate-blog	Tue 9am UTC	Auto-generates blog articles
send-digest	Tuesday 3pm UTC	Sends the weekly newsletter digest
generate-comparisons	Sunday 9am UTC	Pre-generates head-to-head server comparison pages
check-broken-links	Sunday 9am UTC	Checks for broken links across the site
detect-duplicates	Monday 9am UTC	Detects duplicate server entries

Run any bot locally:

npx tsx bots/compute-scores.ts

Scoring

Each server is scored 0-100 across 5 categories:

Category	Weight	Source
Security	30	OSV.dev CVEs, license, authentication, archive status
Maintenance	25	Commit recency, stars, downloads, open issues
Efficiency	20	Token cost of tool schemas in context window
Documentation	15	README quality, tool docs, install configs, examples
Compatibility	10	Transport support (stdio/HTTP), tested clients

See the full methodology at mcppedia.org/methodology.

Contributing

Contributions welcome! Some ways to help:

• Improve server data — edit pages on the site to fix descriptions, add install configs, or categorize servers
• Report issues — open a GitHub issue for bugs or feature requests
• Add features — check open issues for things to work on

License

This project is licensed under the GNU Affero General Public License v3.0.

This means you can freely use, modify, and distribute this code. If you run a modified version as a network service, you must make your source code available under the same license.