
Semantic search ETL pipeline for the MCP server registry. Exposed as both FastAPI and MCP servers
概览
Semantic search API for MCP servers using the official Model Context Protocol registry. Use as: * : https://mcp-registry-search.vercel.app/search?q=kubernetes&limit=2 * : https://mcp-registry-search.vercel.app/servers?limit=10&offset=0 * : https://mcp-registry-search.vercel.app/api/sse Cron job reindexes the entire registry every night. Built with: * mcp-agent cloud * Supabase * Vercel Query the registry using the /search endpoint: https://mcp-registry-search.vercel.app/search?q=kubernetes&limit=2 - q (required): Search query string - limit (optional): Maximum number of results (default: 10) - full_text_weight (optional): Weight for full-text search (default: 1.0) - semantic_weight (optional): Weight for semantic search (default: 1.0) List all servers with pagination using the /servers endpoint:
README
以下为原文 README 节选,完整内容请查看原文。
MCP Registry Search
Semantic search API for MCP servers using the official Model Context Protocol registry.
Use as:
- REST API (search): https://mcp-registry-search.vercel.app/search?q=kubernetes&limit=2
- REST API (list): https://mcp-registry-search.vercel.app/servers?limit=10&offset=0
- MCP Server (SSE): https://mcp-registry-search.vercel.app/api/sse
Cron job reindexes the entire registry every night.
Built with:
- mcp-agent cloud
- Supabase
- Vercel
Usage
Search REST API
Query the registry using the /search endpoint:
Endpoint: https://mcp-registry-search.vercel.app/search?q=kubernetes&limit=2
Example:
curl "https://mcp-registry-search.vercel.app/search?q=kubernetes&limit=2"
Example Response:
{
"results": [
{
"id": 259,
"name": "io.github.vfarcic/dot-ai",
"description": "AI-powered development platform for Kubernetes deployments and intelligent automation",
"version": "0.101.0",
"repository": {
"url": "https://github.com/vfarcic/dot-ai",
"source": "github"
},
"packages": [
{
"version": "0.101.0",
"transport": {
"type": "stdio"
},
"identifier": "@vfarcic/dot-ai",
"registryType": "npm"
}
],
"remotes": [],
"similarity_score": 0.606411385574579
},
{
"id": 272,
"name": "io.github.containers/kubernetes-mcp-server",
"description": "An MCP server that provides [describe what your server does]",
"version": "1.0.0",
"repository": {
"url": "https://github.com/containers/kubernetes-mcp-server",
"source": "github"
},
"packages": [],
"remotes": [],
"similarity_score": 0.451448836663574
}
],
"query": "kubernetes",
"limit": 2,
"count": 2
}
Query Parameters:
q(required): Search query stringlimit(optional): Maximum number of results (default: 10)full_text_weight(optional): Weight for full-text search (default: 1.0)semantic_weight(optional): Weight for semantic search (default: 1.0)
List Servers API
List all servers with pagination using the /servers endpoint:
Endpoint: https://mcp-registry-search.vercel.app/servers?limit=10&offset=0
Example:
curl "https://mcp-registry-search.vercel.app/servers?limit=5"
Query Parameters:
limit(optional): Maximum number of results (default: 100)offset(optional): Number of results to skip (default: 0)
MCP Server
Connect to the MCP server via SSE for direct integration with MCP clients:
Endpoint: https://mcp-registry-search.vercel.app/api/sse
Available Tools:
search_mcp_servers(query, limit, full_text_weight, semantic_weight)- Search servers using hybrid searchlist_mcp_servers(limit, offset)- List all servers with pagination
Add to your MCP client config:
{
"mcpServers": {
安装
This server does not publish a one-line install command.
Open the repository installation guide 新手?先看 mcpskill.org 教程 →配置
{
"mcpServers": {
"registry-search": {
"url": "https://mcp-registry-search.vercel.app/api/sse",
"transport": {
"type": "sse"
}
}
}
}