Back to directory

Yuxi

A self-hosted, multi-tenant knowledge agent platform combining RAG, a knowledge graph, multi-agent orchestration, MCP/Skills, a sandbox and permission management in one workspace.

Yuxi

Yuxi is a self-hosted, multi-tenant knowledge agent platform that lets an organization run custom agents with controlled permissions while accumulating an internal knowledge base. It combines document retrieval (RAG), a knowledge graph, LangGraph-based multi-agent orchestration, MCP and Skills, sandboxed tools, and permission management inside a single workspace, aiming to give every user something like their own cloud-hosted coding agent.

Its Agentic RAG suite uploads documents and handles parsing, chunking and vector indexing; multi-step tasks combine tools, MCP, Skills, sub-agents and a sandbox to produce previewable, downloadable files; and entities and relationships extracted from a Milvus knowledge base's document chunks are written into Neo4j and used in retrieval. Team usage -- knowledge bases, agents, Skills and models -- can be scoped by user, department and sharing range, and retrieval and full agent-task quality can be compared and evaluated, including with Langfuse datasets.

It runs entirely through Docker Compose, with LangGraph, Vue, FastAPI, Milvus, Neo4j, PostgreSQL, MinerU and PaddleOCR as its core services, and an initialization script generates the required environment file and security keys before the first launch.

More in Agent frameworks

See all 902
  • Recurse

    recurse.run

    A serverless harness that lets a coding agent design, test, and deploy its own specialist sub-agents as tools, MCP servers, or bots, defined by a schema-validated manifest.

  • ProofForge

    shulgin.is-a.dev

    An AI-agent pipeline that decomposes a mathematics problem, proves each piece, and formalizes the result in Lean 4 and Mathlib until the Lean kernel accepts it, with six of its proofs already merged into Google DeepMind's formal-conjectures repository.

  • local-coder

    github.com

    Builds a local coding-agent team on top of Ollama and OpenCode by detecting the machine's hardware, recommending models for seven specialist roles, and testing whether each assigned model can actually use tools, edit a repository and pass verification.