Yuxi (语析) is an open-source intelligent knowledge base and knowledge graph agent development platform. It empowers developers to build production-grade AI applications that combine Retrieval-Augmented Generation (RAG) with knowledge graph reasoning — all within a single, integrated system built on LangGraph v1, Vue.js 3, FastAPI, and LightRAG.
Unlike platforms that merely wrap a large language model behind a chat interface, Yuxi weaves together agent orchestration, document understanding, knowledge retrieval, and graph-based reasoning into a runtime where these capabilities genuinely collaborate. Whether you're building an enterprise knowledge assistant, a research report generator, or a multi-agent workflow, Yuxi provides the scaffolding so you can focus on domain logic rather than infrastructure plumbing.
Sources: README.md, docs/intro/project-overview.md
Core Features at a Glance
Yuxi's design philosophy centers on three pillars: configurable agents, knowledge-backed reasoning, and production-ready engineering. The table below summarizes what each pillar delivers.
| Capability | Description | Key Technologies |
|---|---|---|
| Agent Development | Build and configure conversational agents with tools, Skills, MCPs, SubAgents, and middleware | LangGraph v1, BaseAgent class, toolkits |
| Knowledge Base (RAG) | End-to-end document pipeline: upload → parse → chunk → embed → retrieve → evaluate | LightRAG, Milvus, Embedding/Rerank models |
