What is WeKnora used for?

WeKnora is a Tencent-backed LLM knowledge platform that turns documents into searchable RAG, a reasoning agent and a self-maintaining wiki, with tool use and MCP support. It is a good fit for enterprise document question answering, building an AI knowledge base from raw files and agents that act on internal knowledge.

Is WeKnora open source?

Yes. WeKnora is open source under a custom licence. Its source code is on GitHub at Tencent/WeKnora and is written mainly in Go.

Yes. WeKnora is open source, so the software itself is free to use under the terms of its own licence.

Can I self-host WeKnora?

Yes. WeKnora can be self-hosted on your own server or infrastructure.

What are some alternatives to WeKnora?

Similar open-source tools in the AI Infrastructure category include Langflow, RAGFlow and Yuxi. SaaS products in the same category include Cohere, Unstructured and AI Stats.

Is WeKnora actively maintained?

Yes. The most recent commit to WeKnora was on 1 October 2026, and the latest release is v0.8.2, published on 24 September 2026. The project has 32k stars on GitHub.

WeKnora: open-source AI infrastructure

About WeKnora

WeKnora is an open-source framework that uses large language models to understand enterprise documents, retrieve them semantically and reason over their contents. It brings a team's documents together to be searched, analyzed and kept up to date, and it offers three ways to work with the same knowledge bases: retrieval-augmented generation for lookups, an agent for multi-step tasks, and a wiki for organizing knowledge.

The agent has an extensive toolbox. Skills installed from ClawHub, SkillHub, Git or ZIP files run inside persistent sandboxes using Docker, E2B or Cube, with an interactive terminal and a graphical desktop next to the chat. A browser extension allows the agent to drive the user's own Chrome or Edge, while external MCP services, even ones requiring OAuth, can be attached and switched on one tool at a time. Cross-session memory keeps profile and preferences, and topics cover embeddings, reranking, Ollama and OpenAI support for question answering.

WeKnora comes from Tencent and is written in Go. Its license is listed as 'Other' on GitHub, so review the repository terms before adopting it. Documentation and a quick start are provided in English, Simplified Chinese, Japanese and Korean.

Key features

Retrieval-augmented generation over team documents
Reasoning agent for multi-step tasks
Self-maintaining wiki of knowledge
Sandboxed skills in Docker, E2B or Cube
Browser control through a Chrome or Edge extension
MCP service connections with OAuth
Cross-session long-term memory

Good fit for

→Enterprise document question answering
→Building an AI knowledge base from raw files
→Agents that act on internal knowledge

Built with: Go
Tags: rag; llm; knowledge-base; ai-agent; mcp; go; ollama; document-qa; tencent

WeKnora: questions and answers

What is WeKnora used for?: WeKnora is a Tencent-backed LLM knowledge platform that turns documents into searchable RAG, a reasoning agent and a self-maintaining wiki, with tool use and MCP support. It is a good fit for enterprise document question answering, building an AI knowledge base from raw files and agents that act on internal knowledge.
Is WeKnora open source?: Yes. WeKnora is open source under a custom licence. Its source code is on GitHub at Tencent/WeKnora and is written mainly in Go.
Is WeKnora free?: Yes. WeKnora is open source, so the software itself is free to use under the terms of its own licence.
Can I self-host WeKnora?: Yes. WeKnora can be self-hosted on your own server or infrastructure.
What are some alternatives to WeKnora?: Similar open-source tools in the AI Infrastructure category include Langflow, RAGFlow and Yuxi. SaaS products in the same category include Cohere, Unstructured and AI Stats.
Is WeKnora actively maintained?: Yes. The most recent commit to WeKnora was on 1 October 2026, and the latest release is v0.8.2, published on 24 September 2026. The project has 32k stars on GitHub.