Wiki is Wiki Management for AI Applications: the operating system for what your AI is allowed to know, cite, and say. Most AI failures are not model failures. They are knowledge failures. Unreviewed PDFs in a vector store. Internal policy mixed into a public chatbot. Stale refund rules still retrieved as truth. Wiki exists so teams can govern knowledge the way they already govern code and access. With Wiki you: - Collect documents into owned collections with lifecycle status, source links, and version history. - Verify sources through review queues and approval workflows so draft content stays out of production answers. - Define AI Applications, audiences, and access policies so each assistant retrieves only the collections it is allowed to use. - Test questions in the Answer Playground with citations, coverage, refusals, and evaluation runs before customers see answers. - Publish Docs Sites from the same governed source, including custom domains and Docs Assistant on higher plans. - Monitor Knowledge Health: stale sources, conflicts, review queues, and policy health before AI relies on them. - Connect your own bot over HTTP with Wiki Connect for policy-constrained, citation-backed answers. Wiki is not a traditional knowledge base and not a thin RAG wrapper. It is the control plane between your sources and your AI applications: public-safe collections for customer-facing apps, internal collections for employee copilots, and proof that an answer came from something someone reviewed. Start with one source, one policy, and one answer worth trusting. Plans from $29/month with a $1 14-day trial.