WeKnora
New Active Docs only Last checked 2026-11-16 Permissive (MIT/Apache/BSD) Server / Docker
๐ฌ One weekly email: local RAG releases, hardware compatibility updates, and practical setup guides. Free. Unsubscribe anytime.
Updates on X: @mkjerman
What it is
WeKnora is Tencent's open-source knowledge framework for team document Q&A. It combines three modes on the same knowledge bases: RAG search with citations, an agent for multi-step tasks (with sandboxed skills and a built-in MCP server), and an auto-generated wiki with a knowledge graph.
Self-hosting is Docker Compose (web UI on localhost) with swappable LLMs, vector DBs and storage; the README documents running local models through Ollama. It also has multi-workspace RBAC and connectors (Notion, Confluence, GitLab, Feishu, RSS). There's a hosted WeChat option too, but the self-hosted stack keeps data in your environment. RAGShelf has not run it.
- License
- MIT (third-party components excepted)
- Latest release
- v0.9.0
- Release date
- 2026-10-29
- Hardware
- Docker Compose or Kubernetes server; optional Neo4j, MinIO and Langfuse profiles unverified
- Document types
- 10+ formats incl. PDF, Word, Excel, images, XMind; sync from Notion, Confluence, GitLab, Feishu, RSS and more unverified
- OCR / tables / retrieval
- Multimodal parsing, hybrid search and citations per README; knowledge graph via Neo4j profile unverified
- Cloud calls
- None required when self-hosted with Ollama models; 29 built-in model vendors (OpenAI, DeepSeek, Qwen, Gemini, etc.) optional unverified
- Fully offline
- Yes
- GitHub stars
- 37,542
- Last commit
- 2026-11-16 (last push, GitHub API 2026-11-16)
Install
Copied from the official README. Check the project docs for options.
git clone https://github.com/Tencent/WeKnora.gitcd WeKnora && cp .env.example .envdocker compose pull && docker compose up -dRetrieval
Hybrid (vector + keyword) Agentic / iterative Graph RAG
Based on: README: hybrid search with citations; agent mode; wiki knowledge graph / Neo4j profile (checked 2026-11-09) (checked 2026-10-30). Vector vs. vectorless RAG explained โ
NVIDIA DGX Spark
DGX Spark: Unknown
No DGX Spark report found yet. We only mark a tool as working when there is a published source.
Last updated: 2026-11-16. License and release data from the GitHub API. Fields marked unverified have not yet been checked against official docs. How we know: Docs only (see the legend). Spot something wrong? Send a correction.
๐ฌ Get next week's tool
One new local RAG tool in your inbox every Monday. Free, unsubscribe anytime.
Sample issue: "AirLLM: run a 70B model on a 4GB GPU by loading one layer at a time. What it's good for (slow, private batch Q&A over your docs), what it isn't (fast chat), and the install command." Plus one new setup recipe or page, like our local vs. cloud cost calculator.