Search documents, run knowledge tasks, and maintain a wiki

Q: How do I install WeKnora?

Run dsh plugin --profile web add @wxg-prc-cpg/dsh-weknora in a terminal. The --profile value names the profile the plugin is added to, so change it if you run a different one.

At a glance

RAG with hybrid retrieval, multimodal document parsing, and citations

Agent tasks with skills, persistent sandboxes, and external MCP tools

Wiki organization with a knowledge graph and rollback

Automatic synchronization from sources including Notion, Confluence, GitLab, and Feishu

Local or private-cloud deployment with configurable models, vector databases, and storage

What does WeKnora do?

WeKnora is an open-source knowledge platform for document understanding, retrieval, and reasoning. It gives teams three ways to work with the same knowledge bases: retrieve answers through RAG, carry out multi-step agent tasks, and organize material into a wiki.

For DSH users, the README lists the npm package @wxg-prc-cpg/dsh-weknora. The platform also exposes a built-in MCP server for access from AI tools.

Core features

Retrieve information with hybrid search and citations, using document parsing that covers PDF, Word, images, Excel, XMind, and other formats.
Run agent skills in session-persistent Docker, E2B, or Cube sandboxes, and connect external MCP services with tools enabled individually.
Organize documents into a wiki with a knowledge graph and rollback support.
Synchronize knowledge from services such as Feishu, Confluence, GitLab, Notion, Yuque, and RSS. Uploaded folders retain their directory structure.
Manage access through workspace roles, resource ownership, and audit logs, with Langfuse tracing for agent steps, token usage, and pipelines.

How to use WeKnora

Choose a deployment route: the online WeChat Dialog platform, a Tencent Cloud Lighthouse template, or a self-hosted Docker or Kubernetes environment. For self-hosting, configure models, storage, and networking, then add documents or connect supported data sources.

Use RAG for document questions, the agent for multi-step tasks, and the wiki for organizing knowledge.

The README identifies the DSH package @wxg-prc-cpg/dsh-weknora, but the supplied excerpt does not include its connection settings or usage commands. Consult the package documentation for those details.

Questions about WeKnora

How do I install WeKnora?

Run dsh plugin --profile web add @wxg-prc-cpg/dsh-weknora in a terminal. The --profile value names the profile the plugin is added to, so change it if you run a different one.

Is WeKnora free to use?

The source is public at Tencent/WeKnora, but the repository does not state a license. Without one you have no explicit right to reuse the code, so ask the author before building on it.

Is WeKnora still maintained?

The last code push was on Nov 9, 2026, and the latest release is v0.9.0. We last refreshed this data on Nov 12, 2026. DeepSeek Harness changes quickly, so a plugin untouched for months may no longer load.

Has Plugin DSH tested WeKnora?

No. It is listed from its public repository and we have not installed it. Read the source and recent issues before you run it.

How is WeKnora relevant to DSH?

The README identifies the npm package @wxg-prc-cpg/dsh-weknora. It also describes a built-in MCP server for AI tools. The supplied excerpt does not specify the DSH package's commands or configuration.

Do RAG, agents, and the wiki use separate knowledge bases?

All three modes operate on the same knowledge bases, supporting retrieval, multi-step work, and knowledge organization.

Can I run it in my own environment?

The README supports local and private-cloud deployment, including Docker and Kubernetes. Models, vector databases, and storage backends can be selected independently.