Beijing QZ Technology Knowledge Base
AI-Native Knowledge

Make team knowledge
ask-able, findable, visual

An AI-powered enterprise knowledge base: ask in natural language and get streamed answers with citations, find anything with hybrid semantic retrieval, and explore the auto-built knowledge graph.

48852Published docs
349115Graph entities
257797Relations

Built for how teams actually work

From writing to retrieval to reasoning — every layer is AI-assisted.

AI semantic Q&A

Streamed answers grounded only in your knowledge base, with source citations — no hallucinations, multi-turn context.

Hybrid retrieval + rerank

Vector (bge-m3) and keyword recall, fused and re-ranked by a cross-encoder for pinpoint relevance.

Follow-up rewriting

Pronouns and ellipsis in follow-ups are auto-resolved into standalone queries, so context never breaks retrieval.

Entity knowledge graph

Entities and relations are extracted automatically and rendered as an interactive force-directed graph.

Mermaid visual editor

Code on the left, canvas on the right — drag to draw flowcharts and architecture diagrams, synced both ways.

Personal knowledge base

A private space for every member, with apply-to-publish review, collaboration and version history.

Powered by a modern RAG stack

Open, swappable, OpenAI-compatible.

pgvector bge-m3 embeddings bge-reranker rerank query rewrite chunk + overlap GraphRAG Mermaid Markdown versioning

Start exploring your knowledge

Browse the library or just ask the AI a question.