Knowledge Base Search
Semantic RAG search across all beckmann.ai posts — returns the most relevant articles incl. excerpts as grounding for answers.
What does the tool do?
Makes the entire beckmann.ai knowledge base searchable from your AI client: the tool embeds your question, ranks all blog posts by vector similarity and returns hits with title, URL, section, relevance score, teaser and the best-matching excerpt.
Your assistant can answer with sources and link straight to them — instead of guessing. If semantic search is unavailable, a lexical fallback kicks in automatically.
- Semantic search (embeddings) instead of bare keyword matching
- Hits include an excerpt — ideal as RAG grounding
- Searches both languages, optionally limited to DE or EN
- Lexical fallback — search never breaks
How to talk to it
Example prompts to copy — your assistant then calls the tool on its own.
“Search the Beckmann knowledge base for “EU AI Act obligations for companies” and summarize the hits.”
“What does beckmann.ai write about MCP servers? Use search_knowledge and link the sources.”
Parameters
| Name | Required | Description |
|---|---|---|
query | yes | Question or keywords (2–200 characters). |
language | optional | “de” or “en” — default: both languages. |
limit | optional | Maximum number of hits 1–10 (default 5). |