Early developmentNot everything described here has shipped yet.

All use cases

Enterprise knowledge

Knowledge your teams and AI can trust

Enterprise knowledge is scattered, unsourced, and ungoverned, and now you're pointing AI at it. An assistant is only as trustworthy as what it reads. NeoWiki makes your knowledge structured, governed, and sourced: one source your teams can act on and your AI can be trusted to build on.

Retention policystatus approvedreviewed Mar 2026LegalCustomer recordsowned byapplies toaudits

Your AI is only as good as what it reads

Your teams are already pointing assistants and copilots at the company's knowledge, and agents won't be far behind. The answers come back fluent and fast, and some of them are wrong: a policy superseded last quarter, an owner who left, a number no one can source. The model isn't the weak link; the knowledge under it is. Most of it is a sprawl of documents and tools, unstructured and ungoverned, with no way to tell current from stale or asserted from sourced.

The problem was never getting an answer. It is trusting the answer, and being able to stand behind it when someone asks where it came from. AI raises the stakes on ungoverned knowledge; it does not fix them.

What NeoWiki commits to

NeoWiki makes that knowledge trustworthy by construction. Schemas turn the parts that matter (owners, statuses, dates, dependencies, classifications) into typed, validated fields, and the prose that explains them stays on the same record.

Every change is versioned and attributed, review and approval workflows apply, and the governance metadata itself (approval state, last reviewer, review date) is structured data you can query. "Every control approved this quarter" or "every runbook not reviewed in a year" is a question with a real answer, whether an auditor or an agent is asking.

Ask across the whole graph: what depends on what, what is overdue for review. Answers are computed from the current records.

And because the knowledge is structured, governed, and sourced, it is finally a foundation an assistant can stand on. Point your RAG, copilots, or agents at NeoWiki and what they retrieve carries a real owner, a real approval, and a real source; GraphRAG pulls connected subgraphs rather than loose snippets of text. When an agent writes back, it goes through the same review and attribution as a human edit, so nothing lands unchecked. Bring your own AI, and ground it on knowledge that holds up.

It is open source and exportable, so the knowledge, and the AI you build on it, stays yours.

Proven, and yours to run

Hallo Welt!, the team behind BlueSpice, the leading commercial MediaWiki distribution, chose NeoWiki as the structured-data layer of its next major version, BlueSpice Galaxy. They build enterprise knowledge tooling for a living.

Run it the way that fits. NeoWiki is open source: host it yourself, or run it on our managed hosting. If you want the wider BlueSpice distribution on top, that is a path too, with the same structured core underneath.

Where NeoWiki is going

The enterprise capabilities are landing now: richer query results on the page, more property types, per-record access control, deeper approval and review tooling, and an MCP endpoint agents connect through. It is open source as it ships, for any team to run.

Want the full picture? See all of NeoWiki's capabilities.

Recognize your situation?

Let's talk through the specifics.