lunoo-mcp

MCP serverDev tools

Rankings gives your AI access to consensus rankings covering 75,000+ products across 2,600+ niches. Once added, your AI can look up how products stack up in a niche and point you to the top options without you sorting through scattered sources.

Use lunoo-mcp in Claude, ChatGPT or Ahel Desktop

Free. Sign in, add lunoo-mcp and connect your AI. About a minute.

Also: Claude Code · Cursor · Codex

Then ask your AI: use lunoo-mcp to find niche

Details

Available today. Use it from your connected AI after setup.

After adding Rankings, ask your AI to look up rankings for a product or niche you have in mind, such as the top options in a category you are researching.

lunoo-mcpStart free

What your AI can do with it

  • Look up the consensus ranking for any of 75,000+ products
  • Find the top-ranked products in a specific niche
  • Compare how products stack up within one of 2,600+ niches
  • Check where a given product stands among its competitors
  • Explore rankings across many niches to spot leading options

Install lunoo-mcp

The server’s own address, for the clients that take one directly. Or connect Ahel once and every client you use reads it from one address, with the account kept on Ahel rather than in each client’s config.

  • Claude Code

    claude mcp add --transport http --scope user lunoo-mcp 'https://lunoo.com/mcp'

    Run it once in your project, then open /mcp to approve any sign-in the server asks for.

  • Claude Desktop

    https://lunoo.com/mcp

    Add a custom connector in Settings, paste this address, and approve the sign-in.

  • Cursor

    cursor://anysphere.cursor-deeplink/mcp/install?name=lunoo-mcp&config=eyJ1cmwiOiJodHRwczovL2x1bm9vLmNvbS9tY3AifQ==

    Open the link and Cursor adds the server at that address.

  • ChatGPT

    https://lunoo.com/mcp

    In Settings, enable Developer mode, create an MCP app, and paste this address. Your plan and workspace must allow custom apps.

  • Codex

    codex mcp add lunoo-mcp --url 'https://lunoo.com/mcp'

    Run it once, then sign in with codex mcp login lunoo-mcp if the server asks for an account.

From the project's README

As published by peakproductivity/lunoo-mcp in README.md.

Public Model Context Protocol (MCP) server for Lunoo, a rankings platform covering more than 231,000 items across more than 2,600 categories on a 0-10 scale.

Ask for rankings in normal language, from anime and laptops to software, travel, books, and thousands of other categories. The server resolves the closest live category, returns scored items with Lunoo links, and refuses unsupported queries instead of inventing a result.

Overview

This is a read-only public proxy over Lunoo's public web API. No authentication required, no database access, no API keys, no cost.

  • Live endpoint: https://lunoo.com/mcp (Streamable HTTP, JSON-RPC 2.0)
  • Transport: Stateless POST over HTTP
  • Caching: Browse page 24h, rankings and items 15 minutes
  • Timeout: 10 seconds per upstream request
  • Coverage: 2,677+ categories and 231,982+ ranked items, verified on 2026-08-21
  • Privacy: No raw prompts, tool arguments, or tool results are stored. See https://lunoo.com/privacy
  • Safety: All three tools are read-only and expose MCP read-only annotations

Security and abuse controls

  • The server listens on 127.0.0.1 by default so a reverse proxy can expose only the intended HTTPS route.
  • Each client is limited to 120 MCP requests per minute.
  • No more than 20 MCP requests and 12 upstream Lunoo requests run at once.
  • JSON bodies are limited to 32 KB, JSON-RPC batches to 20 messages, and tool inputs to documented lengths.
  • The in-memory cache is limited to 256 entries and evicts old keys.
  • User input is encoded into fixed Lunoo paths. It cannot select an arbitrary upstream host.
  • The server does not execute commands, evaluate user code, read local files, or expose environment variables.

Self-hosted deployments should keep the Node port private and expose /mcp through an HTTPS reverse proxy. Do not publish port 8972 directly.

Quick Start

Claude custom connector

Open Claude's connector settings, choose Add custom connector, and enter:

Name: Lunoo Rankings
URL: https://lunoo.com/mcp

No authentication is required. Test with a request such as Find the best laptop for video editing using Lunoo.

Claude Code

In your project's .claude/settings.json:

{
  "mcpServers": {
    "lunoo": {
      "url": "https://lunoo.com/mcp"
    }
  }
}

Or run from any Claude Code session:

claude mcp add --transport http lunoo https://lunoo.com/mcp

Generic HTTP Client

Post JSON-RPC 2.0 requests to https://lunoo.com/mcp. Example:

curl -X POST https://lunoo.com/mcp \
  -H 'Content-Type: application/json' \
  -d '{
    "jsonrpc": "2.0",
    "method": "tools/list",
    "id": 1
  }'

Local Development (self-hosted)

npm install
npm start
# Server listens on http://localhost:8972

Health check:

curl http://localhost:8972/healthz
# {"ok":true,"service":"lunoo-mcp"}

Tools

find_niche

Resolve a normal-language request to live ranking categories. Returns up to 10 matches with confidence, matched terms, item counts, and tracked Lunoo links.

Input:

{
  "query": "string - normal-language request, e.g. 'best laptop for video editing'"
}

Output: Array of niches with slug, display name, Lunoo URL, and JSON endpoint URL.

Example Request:

{
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "name": "find_niche",
    "arguments": { "query": "anime" }
  },
  "id": 1
}

Example Response:

{
  "jsonrpc": "2.0",
  "result": {
    "isError": false,
    "content": [
      {
        "type": "text",
        "text": "[\n  {\n    \"slug\": \"anime\",\n    \"name\": \"anime\",\n    \"url\": \"https://lunoo.com/best/anime\",\n    \"rankings_json\": \"https://lunoo.com/best/anime.json\"\n  },\n  ...\n]"
      }
    ]
  },
  "id": 1
}

get_rankings

Fetch ranked items for a category slug or a normal-language request. Returns the resolved category, match explanation, score metadata, and top items.

Input:

{
  "niche": "string - niche slug, e.g. 'anime'",
  "limit": "number? - items to return, 1-50, default 10"
}

Output: Niche metadata plus ranked items with rank, name, score (0-10), description, URL, and confidence label.

Example Request:

{
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "name": "get_rankings",
    "arguments": { "niche": "anime", "limit": 3 }
  },
  "id": 2
}

Example Response:

{
  "jsonrpc": "2.0",
  "result": {
    "isError": false,
    "content": [
      {
        "type": "text",
        "text": "{\n  \"name\": \"anime\",\n  \"url\": \"https://lunoo.com/best/anime\",\n  \"item_count\": 1847,\n  \"updated_at\": \"2026-07-02T08:15:00Z\",\n  \"score_scale\": \"0-10\",\n  \"ranking_method\": \"Consensus-based scoring\",\n  \"top_items\": [\n    {\n      \"rank\": 1,\n      \"name\": \"Frieren: Beyond Journey's End\",\n      \"score\": 9.15,\n      \"url\": \"https://lunoo.com/item/frieren-beyond-journeys-end\",\n      \"description\": \"A fantasy adventure about a thousand-year journey...\"\n    },\n    ...\n  ]\n}"
      }
    ]
  },
  "id": 2
}

get_item

Extract structured metadata (JSON-LD) for an item by slug. Parses all schema.org blocks embedded on the item page.

Input:

{
  "slug": "string - item slug, e.g. 'frieren-beyond-journeys-end'"
}

Output: Item URL and array of parsed JSON-LD objects (BreadcrumbList, AggregateRating, etc.).

Example Request:

{
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "name": "get_item",
    "arguments": { "slug": "frieren-beyond-journeys-end" }
  },
  "id": 3
}

Example Response:

{
  "jsonrpc": "2.0",
  "result": {
    "isError": false,
    "content": [
      {
        "type": "text",
        "text": "{\n  \"url\": \"https://lunoo.com/item/frieren-beyond-journeys-end\",\n  \"structured_data\": [\n    {\n      \"@context\": \"https://schema.org\",\n      \"@type\": \"AggregateRating\",\n      \"ratingValue\": 9.15,\n      \"ratingCount\": 847\n    },\n    ...\n  ]\n}"
      }
    ]
  },
  "id": 3
}

MCP Initialization

The server implements Model Context Protocol v2025-03-26. Initialize with:

{
  "jsonrpc": "2.0",
  "method": "initialize",
  "params": {
    "protocolVersion": "2025-03-26",
    "capabilities": {},
    "clientInfo": {
      "name": "my-client",
      "version": "1.0.0"
    }
  },
  "id": 0
}

Response includes server capabilities, protocol version, and instructions.

Caching Behavior

Results are cached locally in the live service at https://lunoo.com/mcp:

  • Browse page (used by find_niche): 24 hours
  • Rankings JSON (used by get_rankings): 15 minutes
  • Item pages (used by get_item): 15 minutes

When self-hosting, cache is in-memory per server instance and cleared on restart.

Error Handling

On failure (404, timeout, network error), tools return:

{
  "isError": true,
  "content": [
    {
      "type": "text",
      "text": "Niche \"invalid-slug\" not found (404). Try find_niche to discover available niches."
    }
  ]
}

Never throws unhandled exceptions; all errors are JSON-RPC compliant.

Citation

Link to full rankings pages on Lunoo as:

https://lunoo.com/best/{slug}

Example: https://lunoo.com/best/anime

See https://lunoo.com/llms.txt for how to cite Lunoo rankings in LLM outputs.

Self-Hosted Deployment

Prerequisites

Node 18+, npm

Installation

git clone https://github.com/PeakProductivity/lunoo-mcp.git
cd lunoo-mcp
npm install

Running

npm start
# or with custom port:
PORT=9000 npm start

Production Setup

  1. Use a process manager (PM2, systemd, etc.):
pm2 start server.mjs --name lunoo-mcp
  1. Reverse proxy (nginx):
location /mcp {
  proxy_pass http://localhost:8972;
  proxy_set_header Host $host;
  proxy_set_header X-Real-IP $remote_addr;
}
  1. Monitor cache size and logs periodically.

CORS Policy

The live service accepts browser requests only from the configured Claude origins. Requests from arbitrary browser origins are rejected. Server-side MCP clients without an Origin header are accepted.

Self-hosted instances use the same safe defaults. Override the browser allowlist with the comma-separated MCP_ALLOWED_ORIGINS environment variable when necessary.

API Endpoints

MethodPathDescription
POST/mcpJSON-RPC 2.0 endpoint
GET/healthzHealth check

Support

Found an issue with a ranking? Report it on Lunoo: https://lunoo.com

License

MIT. See LICENSE file.

Links

Tools it offers (3)

What this server listed when Ahel dialed its public endpoint in Oct 2026, with no key and no account of yours. The names are the server’s own.

  • find_niche
  • get_rankings
  • get_item

Signals

Last commit
Sep 2026
Advanced
Delivery
rankings MCP server → your Ahel connector (mcp.ahel.ai) → your AI.
Item type
mcp-server
Key
com-lunoo-rankings
Source
github.com/peakproductivity/lunoo-mcp
Hosted endpoint
https://lunoo.com/mcp