WeKnora Tencent/WeKnora

Self-hosted LLM knowledge platform from Tencent that turns your documents into cited RAG Q&A, a reasoning agent and an auto-maintained wiki.

Creator kit + Compare GitHub ↗ Install guide (official) ↗

WeKnora at a glance

Checked 16 Oct 2026

WeKnora is a Go-based knowledge platform deployed with Docker Compose, Helm or a single-binary Lite mode. It parses PDF, Word, Excel, images and more, answers questions over them with source citations, and adds an agent and wiki mode, team workspaces and IM channels. MIT licence (third-party parts keep theirs). It ships no chat model: connect Ollama or one of 29 API vendors. Built for teams; heavier than a personal chat app.

Cite this pageRepoQuarry. “WeKnora.” Checked 16 Oct 2026. https://repoquarry.dev/projects/weknora

How much of ChatGPT Plus does it replace?

1 of 3 jobs. Replaces chatting with your documents (with citations) across shared knowledge bases; answer quality depends on the model you connect.

Chat assistant Not covered

Chat with your documents Partial replacement

Web search answers Not covered

For chat assistant, web search answers, see other ChatGPT Plus alternatives. You may also need to build an interface, an API server and hosting yourself.

Facts, with sources

Every answer links to where we found it. We read the documentation; we did not install it.

Report a correction

QuestionAnswerSourceChecked

Runs on my machine? Yes, self-hosted via Docker Compose, Kubernetes or a single Lite binary Deploy locally or on a private cloud and keep the data in your own environment. github.com · WeKnora#readme ↗ 1d ago

Code licence MIT (third-party components keep their own licences; GitHub shows NOASSERTION) WeKnora is licensed under the MIT except for the third-party components listed below. github.com · LICENSE ↗ 1d ago

Model weights licence No bundled chat weights; licence depends on the model you connect 29 built-in vendors with a generated model catalog, including OpenAI, DeepSeek, Qwen (Alibaba Cloud), Zhipu, Hunyuan, Gemini, MiniMax, NVIDIA, LiteLLM and Ollama. github.com · WeKnora#readme ↗ 1d ago

Commercial use Yes, MIT code; check third-party component and model licences github.com · LICENSE ↗ 1d ago

Paid services needed? None required with local Ollama; paid model APIs, web search APIs and WeKnora Cloud optional To use a local Ollama model, run `ollama serve > /dev/null 2>&1 &` first. github.com · WeKnora#run-with-docker-compose ↗ 1d ago

Hardware Not in the README; it points to the online installation guide github.com · WeKnora#other-ways-to-deploy ↗ 1d ago

Operating systems Official installation docs describe Docker/Kubernetes servers and a source-built macOS desktop app. Windows/Linux/macOS desktop build targets exist, but the docs warn that release installers are not published github.com · 02-installation.md ↗ today

Setup effort Medium; Docker Compose with several services, then configure models in the web UI Requires Docker, Docker Compose and Git. github.com · WeKnora#run-with-docker-compose ↗ 1d ago