HubLensRAGTencent/WeKnora
// archived 2026-04-23
Tencent

WeKnora

AI#RAG#LLM#Agents#Knowledge Graph#Vector Database
View on GitHub
14,104

// summary

WeKnora is an open-source, LLM-powered framework designed for enterprise-grade document understanding, semantic retrieval, and autonomous reasoning. It features a ReAct agent for complex multi-step tasks and a Wiki mode that distills raw documents into a structured, interlinked knowledge base. The platform supports multi-source data ingestion, various LLM integrations, and flexible deployment options to ensure complete data sovereignty.

// technical analysis

WeKnora is an open-source, modular knowledge framework designed to transform enterprise documents into living, queryable assets through RAG, autonomous ReAct agents, and an automated Wiki system. By providing a highly extensible architecture, it allows users to swap LLMs, vector databases, and storage backends while maintaining full data sovereignty through self-hosted deployments. The project addresses the challenge of fragmented enterprise information by orchestrating multi-source ingestion and complex reasoning tasks, making it a robust solution for teams needing a continuously evolving, interlinked knowledge base.

// key highlights

01
Wiki Mode automatically distills raw documents into structured, interlinked markdown pages with an interactive knowledge graph for better information discovery.
02
ReAct Agent capabilities enable autonomous multi-step reasoning by orchestrating retrieval, web search, and MCP tools to solve complex queries.
03
Comprehensive data source integration supports auto-syncing from platforms like Feishu, Notion, and Yuque, ensuring knowledge bases stay current.
04
Extensive compatibility with over 20 LLM providers and multiple vector database backends allows for flexible, vendor-neutral infrastructure deployment.
05
Integrated Langfuse observability provides deep visibility into agent reasoning, token usage, and pipeline tracing for performance monitoring.
06
Multi-channel IM support allows users to interact with the knowledge framework directly through platforms like Slack, Telegram, WeCom, and Feishu.

// use cases

01
RAG-based intelligent Q&A for enterprise documents
02
Autonomous ReAct agents for multi-step reasoning and tool orchestration
03
Automated Wiki generation and knowledge graph visualization from raw documents

// getting started

To begin using WeKnora, you can deploy the framework locally using Docker or via Kubernetes using the provided Helm charts for enterprise-grade setups. Once deployed, you can access the Web UI to configure your preferred LLM providers, connect data sources like Notion or Feishu, and start building your knowledge bases. For a quicker start, you may also explore the cloud-hosted knowledge assistant service or the WeChat Mini Program for mobile access.