# WeKnora (Tencent's open-source LLM knowledge platform, v0.9.0)

> Updated 2026-10-30 · type: tool · category: productivity-knowledge · status: active · rev 1

WeKnora, from Tencent, turns your docs into a self-hosted RAG agent and living wiki, and v0.9.0 lets those agents split work and run on a schedule.

- Open source: yes (MIT)
- Self-hostable: yes
- Pricing: Free (MIT). Hosted routes via the WeChat Dialog Open Platform and a Tencent Cloud Lighthouse template; no published prices found.
- Best for: Lean teams that want a self-hosted knowledge base over sales, support or product docs, with an agent that answers questions and runs recurring jobs, without a per-seat enterprise search contract.
- Not for: Teams that want a managed SaaS with SSO and support out of the box, or that cannot keep a Docker service off the public internet (the README warns against exposing it).
- Version: 0.9.0
- Last verified: 2026-10-30

- **Canonical:** https://gtmstacker.com/registry/tool/weknora/
- **Source:** [WeKnora v0.9.0 brings multi-agent collaboration and scheduled tasks to open-source AI knowledge platform](https://x.com/Weixin_WeChat/status/2116036376076722264) — Weixin_WeChat
- **Tags:** productivity-knowledge, rag, knowledge-base, ai-agents, multi-agent, scheduled-tasks, self-hosted, tencent, open-source
- **Repository:** https://github.com/Tencent/WeKnora
- **Version:** 0.9.0

## Is WeKnora (Tencent's open-source LLM knowledge platform, v0.9.0) open source?

Yes, WeKnora (Tencent's open-source LLM knowledge platform, v0.9.0) is open source under the MIT license.

## How much does WeKnora (Tencent's open-source LLM knowledge platform, v0.9.0) cost?

Free (MIT). Hosted routes via the WeChat Dialog Open Platform and a Tencent Cloud Lighthouse template; no published prices found. (free).

## Can I self-host WeKnora (Tencent's open-source LLM knowledge platform, v0.9.0)?

Yes, WeKnora (Tencent's open-source LLM knowledge platform, v0.9.0) can be self-hosted (the source is available under the MIT license).

## Alternatives & related

- [Onyx](https://gtmstacker.com/registry/tool/onyx/)
- [AnythingLLM](https://gtmstacker.com/registry/tool/anythingllm/)


---

**WeKnora** (`Tencent/WeKnora`) is an **MIT-licensed** open-source LLM knowledge platform from Tencent (WeChat). It ingests documents and turns them into a queryable **RAG** index, an autonomous **reasoning agent**, and a **self-maintaining wiki**. Release **v0.9.0** (2026-10-29) adds **multi-agent collaboration** ("delegate independent subtasks, follow progress and bring the results together") and **Automations**: one-off or recurring scheduled tasks with execution history, results and approval notifications.

## Provenance

- License **MIT**, from the LICENSE file text, which excludes a listed set of third-party components (for example opencc under Apache-2.0). The GitHub license API reports `NOASSERTION` because of Tencent's wrapper header, not because the license is unknown.
- Identity: official `Tencent` GitHub org, homepage weknora.weixin.qq.com, created 2025-07-22, not archived.
- Activity as of 2026-10-30 (GitHub API): **36,044 stars, 4,898 forks**, latest release **v0.9.0** published 2026-10-29, last push 2026-10-30.
- Install: Docker Compose (UI on localhost, API on :8080) with optional Neo4j, MinIO and Langfuse profiles; Helm/Kubernetes; a Lite single binary on SQLite; a desktop app built from source.
- **Vendor claim:** the v0.9.0 multi-agent and Automations capabilities are described in the README and release notes and were not independently tested.
- Surfaced via the official WeChat account's X post in the 2026-10-30 pull (the post itself returned HTTP 402; text taken from the signals registry copy). Curated and verified in run `2026-10-30-registry-update-01` (step-3 verification, 2026-10-30).

## Why it matters for a GTM stack

Sales and support teams keep answering the same product, pricing and objection questions from scattered docs. WeKnora gives a lean team a self-hosted knowledge base with an agent on top, and the new Automations mean a recurring job (a weekly digest of new docs, a scheduled check of a battlecard folder) can run without a separate scheduler. Keep it inside your network and test answer quality on your own documents before rolling it out to reps.