What is WeKnora?
WeKnora is an open-source, LLM-powered knowledge framework for enterprise document understanding, semantic retrieval and reasoning. WeKnora brings team documents together so users can search them, reason over their contents and keep knowledge current. The project is published in Tencent's GitHub repository and supports deployment within an organization's own environment.
WeKnora combines retrieval-augmented generation, an autonomous reasoning agent and a self-maintaining wiki around shared knowledge bases. These modes address different tasks: looking up information, completing multi-step work and organizing knowledge. Teams can therefore use the same document collection across retrieval, agent interactions and wiki maintenance.
What WeKnora can do
WeKnora combines document ingestion, knowledge retrieval and agent tools with administration features for shared workspaces. Supported inputs include PDF, Word, images, Excel and XMind. Automatic synchronization connects external document sources, while source citations let readers navigate from an answer to the cited passage in the original document.
- WeKnora synchronizes knowledge from Confluence, GitLab, Notion, Feishu and other listed sources.
- WeKnora lets users edit retrieval chunks, compare changes and roll back revisions.
- WeKnora retains user-confirmed profile details, preferences and facts across sessions.
- WeKnora runs installed agent skills in session-persistent Docker, E2B or Cube sandboxes.
- WeKnora connects external MCP services with tool-by-tool enablement.
- WeKnora provides workspace permissions, resource ownership and audit logs.
How WeKnora works
WeKnora offers a Docker Compose setup that requires Docker, Docker Compose and Git. The documented workflow clones the repository, copies the environment example, adjusts configuration, pulls service images and starts the core services. Users then open the local web interface and follow the onboarding guide, with sample data available in the Quickstart.
WeKnora connects knowledge to WeCom, Feishu, Slack, Telegram and website embeds. A built-in MCP Server exposes knowledge bases to Cursor, Claude and other AI tools, while scoped API keys support programmatic access. BrowserSkill connects the agent to the user's Chrome or Edge browser, including hand-off for logins and CAPTCHAs.
Who WeKnora is for
WeKnora is intended for enterprise teams that need document understanding, semantic search and reasoning over shared information. WeKnora also supports technical operators managing private deployments and developers connecting knowledge bases to applications or AI tools. Workspace access controls and multiple storage instances support the administration of separate team environments.
- WeKnora supports internal document Q&A with citations to source passages.
- WeKnora organizes team knowledge through a self-maintaining wiki.
- WeKnora delivers knowledge-base answers through supported messaging channels.
- WeKnora supports multi-step agent tasks using documents, skills and browser tools.
WeKnora pricing and deployment options
WeKnora is available as an open-source project through GitHub, but the supplied page does not state subscription prices or commercial plan details. WeKnora lists an online WeChat Dialog Open Platform option, a Tencent Cloud Lighthouse application template and self-hosted deployment. The page does not specify charges for the online or cloud options.
WeKnora supports Docker Compose for the standard multi-service deployment and Kubernetes through a Helm chart for production clusters. A Lite single binary targets local or low-resource use with SQLite and an in-memory queue. The desktop app uses the Lite runtime, but no installer is published and building from source is required.
WeKnora strengths and limitations
WeKnora gives teams control over deployment location and allows LLMs, vector databases and storage backends to be swapped. Shared knowledge bases connect retrieval, agents and wiki workflows, while revision controls support document curation. These capabilities come with operational responsibilities: the documentation recommends private-network deployment, appropriate firewall rules, access controls and regular security updates.
- WeKnora self-hosting requires configuration of models, storage and networking.
- WeKnora should not be exposed directly to public networks, according to the production guidance.
- WeKnora sandbox commands run as root, as noted in the release information.
WeKnora FAQ
What is WeKnora?
WeKnora is an open-source, LLM-powered knowledge framework for enterprise document understanding, semantic retrieval and reasoning. WeKnora combines RAG, an autonomous reasoning agent and a self-maintaining wiki using the same knowledge bases.
Is WeKnora free?
WeKnora is available as an open-source project through GitHub. The supplied page does not state subscription pricing or charges for its online platform and Tencent Cloud deployment options, so those access options should not be assumed to be free.
Can WeKnora be self-hosted?
WeKnora supports local and private-cloud deployment using Docker Compose, Kubernetes or a Lite single binary. WeKnora allows teams to configure models, storage and networking and keep data in their own environment.
What documents and data sources does WeKnora support?
WeKnora supports formats including PDF, Word, images, Excel and XMind. WeKnora also automatically synchronizes sources including Feishu wiki, Feishu Drive, Confluence, GitLab, Tencent IMA, Notion, Yuque, DingTalk Docs, Seafile and RSS.
Does WeKnora have an API or MCP integration?
WeKnora provides scoped API keys for programmatic access and a built-in MCP Server for Cursor, Claude and other AI tools. WeKnora can also connect external MCP services, including OAuth-enabled services, with individual tools enabled separately.
Can WeKnora control a web browser?
WeKnora agents can operate the user's Chrome or Edge browser through the BrowserSkill extension. WeKnora supports a live task preview, pause and resume controls, and hand-off to the user for logins and CAPTCHAs.
From the community
No reviews yet
Be the first to review this tool. What you write helps the next person choose.
