← Registry

Data Collection

mcp-hunter.com

MCP Hunter is a searchable directory service that lets you find, inspect, test, and rank MCP servers by name, category, registry data, and weekly leaderboard position.

1 endpoint6 known toolsFirst detected September 29, 2026Last detected September 29, 2026

ENDPOINT 1

https://mcp-hunter.com/mcp

No auth detected

MCP server metadata

Name
MCP Hunter
Version
1.0.0
Capabilities
tools
Server instructions

MCP Hunter is a weekly launch board for products that ship an MCP server. Makers submit, a human reviews every submission, and one ranked leaderboard is published per week at a permanent URL. Read the four paragraphs below before reporting anything from these tools, because the data means something narrower than it looks. WHAT A LISTING IS. A product is here because its maker submitted it and a human approved it. This is not a crawl of the MCP ecosystem and the order on a week's board is the votes readers cast that week, not a quality ranking. The board is young and small. If a search comes back empty, that means nothing is listed here, and it is not evidence that nothing exists. Say which of the two you mean. WHAT THE EVIDENCE MEANS. Where a server was reachable over HTTP, we connected to it once, at submission, and recorded what it returned: the tool count, the tool names, the protocol revision it answered with, and one round trip in milliseconds. That record is dated and past tense. We never re-test on a schedule, so a receipt says what happened on one day and says nothing about whether the server answers now. Never describe a listing as live, online, working, monitored, or currently verified. Quote the date. WHAT CARRIES NO EVIDENCE. A product that starts from a local command is listed with the command shown as text and is never executed, so it has no tool count and no response time. A server that refused us carries what happened and nothing else. Roughly one remote MCP server in four answers 401, which is a working server guarding itself rather than a broken one. Where a listing carries no receipt, report the absence instead of filling it. WHAT WE DO NOT DO. No uptime monitoring, no security audit, no safety rating, no code review, and we never run a submitted command. A listing is a dated record of one moment and is not an endorsement, so it is not evidence that a server is safe to install or trustworthy with data. If you are asked whether a server is safe, say that this board does not answer that. CHECKING AN ADDRESS YOURSELF. check_mcp_server runs the same handshake against any endpoint you give it, listed or not, and returns a dated result with a permanent URL you can hand to a person. Use it when you are about to install something and want to know what it actually exposes. It runs on a queue, so call get_check_result with the returned token a few seconds later.

Known tools 6

search_mcp_servers

Search the MCP products listed on MCP Hunter: name, tagline, description and category.

Inferred read-only
get_mcp_server

The full record for one listed product: what its maker says it does, how it is reached, the dated result of our connection test where there is one, what a guarded server published about signing in, what a package registry held for it, and where it placed on its launch week.

Inferred read-only
weekly_leaderboard

The ranked board for one launch week, most voted first.

Inferred read-only
check_mcp_server

Connect to any Streamable HTTP MCP endpoint and record what it returned: whether it answered, how many tools it exposes, their names, the protocol revision it replied with, and one round trip in milliseconds.

Inferred read-only
get_check_result

The finished result of a check started by check_mcp_server: what the endpoint returned, when, and what it exposed.

Inferred read-only
mcp_ecosystem_stats

Aggregate results from calling MCP servers in bulk: what share of the official registry answered, refused, or went nowhere on a dated sweep, plus a smaller hand-picked corpus checked in depth.

Inferred read-only

CONNECT WITH APPROVAL

Client installation

Review this server and its permissions before adding it. Secret placeholders must be set locally.

Codex

~/.codex/config.toml

[mcp_servers.mcp-hunter]
url = "https://mcp-hunter.com/mcp"
enabled = true
Claude Code

.mcp.json

{
  "mcpServers": {
    "mcp-hunter": {
      "type": "http",
      "url": "https://mcp-hunter.com/mcp"
    }
  }
}
Claude Desktop

Settings → Connectors → Add custom connector

Name: mcp-hunter
Remote MCP URL: https://mcp-hunter.com/mcp

Add this remote URL as a custom connector in Claude Desktop. Availability depends on the user plan and workspace policy.

Cursor

.cursor/mcp.json

{
  "mcpServers": {
    "mcp-hunter": {
      "url": "https://mcp-hunter.com/mcp"
    }
  }
}
Visual Studio Code

.vscode/mcp.json

Add to Visual Studio Code
{
  "servers": {
    "mcp-hunter": {
      "type": "http",
      "url": "https://mcp-hunter.com/mcp"
    }
  }
}
Generic MCP

Client-specific MCP configuration

{
  "name": "mcp-hunter",
  "transport": "streamable-http",
  "url": "https://mcp-hunter.com/mcp"
}
MCP Inspector

Run the official MCP Inspector locally and enter the indexed Streamable HTTP endpoint.

TRUST AND VERIFICATION EVIDENCE

Loading Trust v2 evidence…

Checking the associated registrable domain. The BuiltWith key remains server-side.

Indexed

Evidence is source-attributed and does not guarantee that a third-party server is safe. Risk labels are conservative metadata heuristics.