Everything Next Wiki does today — editor, AI, organization, integrations and administration.

Editor

Block editor with a / menu: headings, lists, task lists, quotes, resizable tables, collapsible blocks, callouts (info, tip, warning, danger), page breaks, images and attachments.
Code and Mermaid diagrams right in the text.
Embeds: YouTube, Vimeo, Loom, Figma, Miro, Google Docs/Sheets/Slides, Airtable, Typeform, CodePen and more. The admin chooses which ones are enabled.
Links between documents with @ or [[. Each document shows where it is mentioned (backlinks).
Board mode: swap the page for a free-form drawing board (Excalidraw).
Presentation: show the document as full-screen slides; each heading becomes a slide.
Export to PDF and Word, or print with a header, footer and margins defined by the workspace.

Artificial intelligence

Chat with sources: ask about the workspace, a collection or a document; the answer cites the source documents. Conversations are saved per project.
Attachments in the chat: up to 3 files per message. PDF, Word (.docx), text and Markdown go in as text (scanned PDFs don’t: there is no OCR); PNG, JPG, WebP and GIF images, up to 5 MB, go to the model.
Agents with tools: they search the base, read documents and attachments, use skills and run in the background. New documents they save right away; changes to existing documents come back as a proposal for you to approve. ClickUp, web search, asking mid-task and calling another agent are available if you enable the tool.
Skills: reusable instructions the team writes once and applies in the chat.
Document types: templates with their own starting text and instructions, so the AI drafts stories, specifications, minutes and whatever else the team standardizes.
In the editor: improve the selected passage, generate a diagram from a passage (flowchart, sequence, class, ER) and generate the document’s mind map.
Conflict analysis: points out documents in a collection that contradict each other (through MCP; no button in the UI yet).
Semantic search (optional, with Qdrant): finds documents by meaning, not only by word.
Multiple providers: OpenAI, Google Gemini, OpenRouter, litellm and Ollama (local models). Each workspace chooses its own.

Collections (projects) with a short code, icon and grouping in the sidebar; drag documents and collections to reorganize.
Search across all content, accent-insensitive, with ⌘K (Ctrl+K).
Comments on passages of text.
Versions of every document, with comparison between versions and restore.
Presence: see who is in the document right now and know when someone saved.
Favorites, tags, the team’s own emojis and a trash with restore (the screen shows the 200 most recent).
Public read-only link to share a document with people who don’t have an account.
Import from Markdown, Confluence exports and another Next Wiki.

ClickUp: agents read tasks and cards (including the text of attachments), and a card becomes a story in the collection.
GitLab and GitHub per collection: open MRs/PRs, tags and a changelog generated as a document.
Signed webhooks (HMAC-SHA256) for documents, collections, comments and invitations.
MCP server: Claude, IDEs and other AI clients read and edit the wiki with your API key. See MCP server.
NextBoard (optional module): backlog, Kanban board, sprints with burndown and teams per project.

Workspace roles (Administrador, Gestor, Gestor de IA, Membro — Administrator, Manager, AI manager, Member) and per-collection permissions for people or groups (Owner, Editor, Comentarista, Leitor — Owner, Editor, Commenter, Reader). Administrators and managers read and edit everything, even in a closed collection.
Sign-in by email link, passkey or SSO with Keycloak (OIDC), configurable from the settings screen; restriction by email domain and self-signup.
Accounts: block someone’s access without deleting what they wrote.
Usage statistics: people, projects, skills, agents and models most used.
Brand on the sign-in screen (system name and logo) and error telemetry via OpenTelemetry.

Want to start? Go to Getting started.