knowledgelib-mcp

MCP server for knowledgelib.io — gives AI agents access to structured, cited knowledge units. Pre-verified product comparisons and reference guides with inline source citations, confidence scores, and freshness tracking.

Quick start

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "knowledgelib": {
      "command": "npx",
      "args": ["knowledgelib-mcp"]
    }
  }
}

Other MCP clients

npx knowledgelib-mcp

The server uses stdio transport and works with any MCP-compatible client.

Tools

query_knowledge

Search across all knowledge units by relevance.

| Parameter | Type | Required | Description | |-----------|------|----------|-------------| | query | string | yes | Search query (e.g., "best wireless earbuds under 150") | | domain | string | no | Filter by domain (e.g., "consumer_electronics", "computing") | | limit | number | no | Max results, 1-20 (default: 3) |

get_unit

Retrieve a specific knowledge unit as raw markdown with YAML frontmatter.

| Parameter | Type | Required | Description | |-----------|------|----------|-------------| | unit_id | string | yes | Unit ID (e.g., "consumer-electronics/audio/wireless-earbuds-under-150/2026") |

Returns the full markdown with frontmatter, comparison tables, use-case recommendations, and inline citations.

list_domains

List all available knowledge domains with unit counts. No parameters. Use this to discover what topics are covered before querying.

Environment variables

| Variable | Default | Description | |----------|---------|-------------| | KNOWLEDGELIB_URL | https://knowledgelib.io | API base URL | | KNOWLEDGELIB_API_KEY | (none) | API key (optional for free tier) |

What you get

Each knowledge unit includes:

• Comparison tables with 8-12 products and specs
• Use-case recommendations (best overall, best budget, best for specific needs)
• Inline citations from 5-8 authoritative sources (RTINGS, Wirecutter, Tom's Guide, etc.)
• Confidence scores (0.70-0.95) based on source agreement
• Freshness tracking with last-verified dates

Why use this instead of searching?

An AI agent answering "what are the best wireless earbuds?" typically runs 3-5 web searches, fetches multiple pages, and spends 5,000-50,000 tokens synthesizing an answer. A knowledgelib unit delivers the same answer in ~600-1,800 tokens, pre-verified and pre-cited.

REST API

The MCP server wraps the knowledgelib.io REST API. You can also use the API directly:

GET https://knowledgelib.io/api/v1/query?q=best+earbuds
GET https://knowledgelib.io/api/v1/units/{id}.md
GET https://knowledgelib.io/api/v1/units/{id}.json
GET https://knowledgelib.io/api/v1/openapi.json

Full API docs: knowledgelib.io/api

License

CC BY-SA 4.0