discover
Find the best MCP server or AI agent for a task. Returns top recommendations with reasoning.
LABSCO SUMMARY
Finding the right server, and checking it, in the same pass.
Search by the job you are trying to do, and every result arrives with the install command, a trust score and the count of known advisories in its dependencies — so the vetting happens before the config entry, not after.
What it is
A search front end for KnyazevAI, an index of 24,500+ MCP servers and AI agents drawn from eight registries. Results come back as structured data — names, descriptions, trust scores, install commands, stars, licences and vulnerability counts.
What you get
- Semantic search by task description rather than by keyword, ranked, with install commands attached to each result
- Browsing with category filters and sorting when you want the field rather than one answer
- The full record for a server: the tools it exposes, its trust breakdown, framework compatibility and packages
- Known CVE and GHSA advisories for a server's dependencies, with severity and affected package
- A recommendation pass across both MCP servers and A2A agents, with the reasoning behind each pick
- Catalog-level statistics: totals by registry, category breakdown and health metrics
Requirements
npx -y knyazevai-mcp-catalog, one entry in your client config. No account and no key. The index is served from knyazevai.work, so it needs network access.
Setup effort
One command — npx -y knyazevai-mcp-catalog
USE CASES
01
Finding a server for a job you cannot name
Describe the task instead of guessing keywords; the ranked results come back with install commands already attached.
02
Checking a server is safe before adding it
Read the known CVEs in its dependencies alongside its trust breakdown, rather than finding out after it is in your config.
03
Comparing the options in a category
Browse with filters and sorting, then pull the full record for the two or three worth a closer look.
04
Seeing what a server actually exposes
Its tool list, framework compatibility and package details — before the decision rather than after the first connection.
05
Getting the shape of the ecosystem
Totals by registry and by category, which is the context for whether a category is crowded or empty.
6 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
get_catalog_stats
Get aggregate statistics: total servers, agents, categories, sources, and health metrics.
get_server
Get full details for a specific MCP server including tools, trust breakdown, install commands, and framework compatibility.
get_server_vulnerabilities
Check known CVE/GHSA vulnerabilities for an MCP server's dependencies.
list_servers
Browse MCP servers with filtering and sorting. Use for category browsing or top-N queries.
search_servers
Semantic search across 24,500+ MCP servers and AI agents by task description. Returns ranked results with trust scores.
BEFORE YOU INSTALL
What to know first
Tools you get
6 tools when we connected.
Requires
npx on PATH
Where it runs
All 4 documented clients run the server on your machine.
CONTEXT COST
Light Standing cost
Rated by the size of the 6 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
TERMINAL
npx -y knyazevai-mcp-catalogMCP Clients
Claude Desktop, Cursor, Windsurf
JSON
{
"mcpServers": {
"mcp-catalog": {
"command": "npx",
"args": [
"-y",
"knyazevai-mcp-catalog"
]
}
}
}VS Code
VS CODE · JSON
{
"servers": {
"mcp-catalog": {
"command": "npx",
"args": [
"-y",
"knyazevai-mcp-catalog"
]
}
}
}
