I’m Hugo, the maker of next-wiki. We built next-wiki after seeing the same pattern everywhere: valuable context gets buried in chats, links, files, incident notes, and AI conversations. Traditional wikis require someone to manually turn all of that into polished documentation. AI chat can answer questions, but often hides the evidence behind the answer. And there are so many Agents having their own memory, rules and prompts which is hard to manage. next-wiki is my attempt to join those these worlds without giving up ownership or review. With next-wiki, you can: • Holds a shared memory store with LLM wiki structure across all agents. • Preserve raw source material separately from Human/AI-generated knowledge • Search and chat over your wiki with visible citations • And ask the native built-in AI to draft, translate, organize, or improve content—while keeping changes as drafts or reviewable proposals • Connect Claude Code, Cursor, OpenCode, and other MCP-compatible clients and let next-wiki be the memory baseline. It’s open source (Apache-2.0) and still early. We’re launching in public because we want to learn from people who are serious about turning scattered context into long-lived, trustworthy knowledge. Try the read-only demo: https://next-wiki.hugogu.cn/ Explore the code: https://github.com/hugogu/next-wiki Thanks for taking a look — I’ll be here all day to answer questions.

Next Wiki

A personal knowledge vault with native AI Agents support

A personal knowledge vault with native AI Agents support