MCP Review – Open-Source MCP Server Directory with Community Ratings

TL;DR

TL;DR: MCP Review is a free, open-source web directory for discovering, rating, and reviewing Model Context Protocol servers for Claude, Cursor, and other AI clients.

Source and Accuracy Notes

⚠️ This section is MANDATORY. All links must be verified from actual source, not guessed.

• Project page: mcpreview.dev ← MUST visit and verify
• Source repository: github.com/ggange/mcp-review ← MUST read README
• License: MIT (verified via GitHub LICENSE file)
• HN launch thread: news.ycombinator.com/item?id=48600911

What Is MCP Review?

MCP Review is an open-source web directory for the Model Context Protocol (MCP) ecosystem. MCP, introduced by Anthropic in late 2024, is an open standard that lets AI assistants like Claude and Cursor connect to external tools, data sources, and services through a standardized server interface.

The directory indexes servers from the official MCP Registry and layers community features on top: ratings, written reviews, per-server tool counts, and official versus community-submitted badges. As of mid-2026, it lists over 1,200 MCP servers.

The site is MIT-licensed and self-hostable. You can submit your own servers, import from GitHub, and vote on reviews. It is built with Next.js 16, Prisma, and PostgreSQL (Neon free tier recommended).

Setup Workflow

Option 1: Use the Live Site

No installation required. Visit mcpreview.dev, browse or search servers, and filter by:

• Official — servers maintained by the tool vendor
• Tool count — number of tools the server exposes
• Category — database, cloud, productivity, etc.

Sign in with GitHub or Google to rate servers or write reviews.

Option 2: Self-Host

Prerequisites

• Node.js 20 or higher
• PostgreSQL database (Neon free tier works)
• Cloudflare R2 for custom icons (optional)

Installation

git clone https://github.com/ggange/mcp-review.git
cd mcp-review
npm install
cp .env.example .env

Configure the required environment variables in .env:

# Database
DATABASE_URL="postgresql://user:host@host:5432/database?sslmode=require"

# NextAuth (optional, for OAuth login)
NEXTAUTH_SECRET="generate with: openssl rand -base64 32"
NEXTAUTH_URL="http://localhost:3000"

# App URL (required in production)
NEXT_PUBLIC_APP_URL="https://your-domain.com"

npm run dev

Visit http://localhost:3000. The site auto-syncs with the official MCP Registry daily, or trigger manually via the cron endpoint.

Key Features

• 1,200+ servers indexed from the official MCP Registry
• Community ratings — trustworthiness and usefulness scores per server
• Written reviews — detailed user reviews with helpful/not-helpful voting
• GitHub import — auto-fill server submission by pasting a repo URL
• Official server badges — distinguishes vendor-maintained from community servers
• Custom icons via R2 — server submitters can upload icons
• GitHub README badges — embed dynamic rating badges in your repo
• Dark mode — full dark theme support
• Responsive — works on desktop and mobile

Practical Evaluation Checklist

• How easy is it to find a server for a specific use case?
  • The site has search and category filtering. Tool-count badges make it easy to spot feature-rich servers.
• Are ratings and reviews helpful?
  • Reviews are community-written and votable. No expert curation, but volume is growing.
• Can I submit my own MCP server?
  • Yes, with GitHub import and tools Markdown file upload for quick population.
• Is it actually open source?
  • Yes, MIT license. Frontend, backend, and database schema are all on GitHub.
• Does it have a self-host option?
  • Yes. Requires Node.js 20+, PostgreSQL, and optional R2 for icons.

Security Notes

• Built with rate limiting, CSRF protection, and input validation (per README)
• OAuth via GitHub or Google — no password auth
• R2 is optional for icons; without it, icons are not uploaded
• Community content (reviews) is user-generated — treat ratings as one data point, not definitive

FAQ

Q: Is MCP Review free to use? A: Yes. The live site at mcpreview.dev is free. Self-hosting is also free under the MIT license. You only pay for hosting and a PostgreSQL database.

Q: How is this different from the official MCP Registry? A: The official registry (github.com/modelcontextprotocol/registry) is a machine-readable list of servers. MCP Review adds a human-readable UI, community ratings, written reviews, icons, and search/filter capabilities.

Q: Can I submit my own MCP server? A: Yes. Log in with GitHub, go to Submit a Server, and either fill the form manually or import from a GitHub repo URL.

Q: What is the license? A: MIT License, verified on GitHub.

Q: Does it work with Cursor or other AI clients besides Claude? A: MCP Review lists servers compatible with any MCP-compatible client, including Cursor, Zed, and other AI coding tools that support the MCP standard.

Conclusion

MCP Review fills the discoverability gap in the rapidly growing MCP ecosystem. The protocol itself is well-documented, but finding the right server for your workflow has been a pain point. A community-driven directory with ratings, reviews, and icons makes discovery much faster. MIT-licensed and straightforward to self-host if you want your own mirror.

If you run MCP servers in production or maintain an AI tool with MCP support, submitting your server to MCP Review and embedding the rating badge in your README is a zero-cost way to get visibility in the developer community.