loading…
Search for a command to run...
loading…
Search 1,500+ pre-verified, cited knowledge units across 16 domains. 6 tools: query, batch query, get unit, list domains, suggest topics, report issues. Confide
Search 1,500+ pre-verified, cited knowledge units across 16 domains. 6 tools: query, batch query, get unit, list domains, suggest topics, report issues. Confidence scores, source provenance, and freshness tracking. Free, no API key required.
AI Knowledge Library — structured, cited knowledge units for AI agents. Pre-verified answers that save tokens, reduce hallucinations, and cite every source.
1,564 knowledge units across 16 domains (consumer electronics, software, business strategy, ERP integration, compliance, energy, finance, and more). Each unit answers one canonical question with:
One API call replaces 5 web searches and 8,000 tokens of parsing.
npx knowledgelib-mcp
Or add to claude_desktop_config.json:
{
"mcpServers": {
"knowledgelib": {
"command": "npx",
"args": ["knowledgelib-mcp"]
}
}
}
POST https://knowledgelib.io/mcp
Streamable HTTP transport, JSON-RPC 2.0, MCP spec 2025-03-26.
# Search
curl https://knowledgelib.io/api/v1/query?q=best+wireless+earbuds+under+150
# Batch search (up to 10 queries)
curl -X POST https://knowledgelib.io/api/v1/batch \
-H "Content-Type: application/json" \
-d '{"queries":[{"q":"earbuds"},{"q":"headphones"}]}'
# Get full unit
curl https://knowledgelib.io/api/v1/units/consumer-electronics/audio/wireless-earbuds-under-150/2026.md
# Health check
curl https://knowledgelib.io/api/v1/health
pip install langchain-knowledgelib
from langchain_knowledgelib import KnowledgelibRetriever
retriever = KnowledgelibRetriever()
docs = retriever.invoke("best wireless earbuds")
npm install n8n-nodes-knowledgelib
| Tool | Description | Read-only |
|---|---|---|
query_knowledge |
Search across all knowledge units with filters | Yes |
batch_query |
Search multiple topics in one call (max 10) | Yes |
get_unit |
Retrieve full markdown content by ID | Yes |
list_domains |
List all domains with unit counts | Yes |
suggest_question |
Submit a topic request for new unit creation | No |
report_issue |
Flag incorrect, outdated, or broken content | No |
All read-only tools are marked with readOnlyHint: true and idempotentHint: true per MCP spec 2025-03-26, enabling parallel execution by agents.
| Type | Count | Description |
|---|---|---|
| product_comparison | 418 | Best-of roundups with decision logic and buy links |
| concept | 336 | Definitions of terms agents often get wrong |
| software_reference | 239 | Code examples, anti-patterns, decision trees |
| execution_recipe | 202 | Step-by-step implementation plans |
| erp_integration | 166 | API capabilities, rate limits, data mapping |
| agent_prompt | 55 | System prompts for pipeline sub-agents |
| assessment | 54 | Structured scoring frameworks |
| decision_framework | 35 | Decision trees with trade-offs |
| benchmark | 28 | Industry benchmarks by segment |
| rule | 28 | Actionable directives with evidence |
CC BY-SA 4.0
Добавь это в claude_desktop_config.json и перезапусти Claude Desktop.
{
"mcpServers": {
"peterbeck111-knowledgelib-io": {
"command": "npx",
"args": []
}
}
}