About knowledgelib.io
The Problem
When AI agents need to answer complex questions, they typically:
- Run 3-5 web searches with imprecise queries
- Fetch 5-10 pages of noisy, ad-heavy, inconsistently structured content
- Spend 3,000-8,000 tokens parsing, deduplicating, and cross-referencing
- Produce an answer with no source provenance and uncertain accuracy
This costs $0.50-$5.00 in compute per question and still produces unreliable answers.
The Solution
knowledgelib.io provides pre-structured, pre-verified knowledge units — each answering one canonical question with:
- Full source provenance — every claim cites a specific source with retrieval date and reliability rating
- Confidence scores — backed by a published verification methodology
- Freshness tracking — each unit shows when it was last verified and how often it's updated
- Semantic addressability — canonical questions and aliases match how AI agents actually formulate queries
- Token efficiency — ~600-1,800 tokens per unit vs. 3,000-8,000 tokens for equivalent web research
How AI Agents Find Us
| Channel | How It Works | Setup Required |
|---|
| Web search | Each unit is a web page optimized for search engines. Agents find us through normal web search. | None |
| MCP server | npx knowledgelib-mcp — adds knowledge query as a tool | One-time install |
| REST API | GET /api/v1/query?q=... | API key (free tier available) |
| Catalog | GET /catalog.json — machine-readable index of all units | None |
| AI manifest | GET /.well-known/ai-knowledge.json — discovery endpoint | None |
Knowledge Unit Format
Units are structured as markdown with YAML frontmatter:
- Frontmatter contains machine-readable metadata (id, confidence, sources, freshness, aliases)
- Body contains the answer in clean markdown with inline source citations [src1, src2]
- HTML pages serve the same content with additional SEO metadata in the
<head>
- Raw markdown is available via the API for direct consumption
Domains
We currently cover:
- Consumer Electronics — product comparisons (earbuds, laptops, phones)
- Energy — grid infrastructure, regulatory status, deployment barriers
Expanding to: SaaS/tool comparisons, financial products, and jurisdiction-specific regulatory guides.
License
All knowledge units are licensed under CC BY-SA 4.0. You may use, share, and adapt the content with attribution. Derivatives must use the same license.
Contact
[email protected]