Aixs Inc. has launched DoXmind, an AI-native document editor aimed squarely at Notion users who want more substantive AI integration than current productivity tools provide. Rather than treating AI as an optional add-on, DoXmind embeds large language model capabilities throughout the core writing workflow — including real-time autocomplete, an AI chat sidebar with full document context, and a knowledge base (RAG) agent that ingests uploaded PDFs and markdown files and returns answers with source citations and relevance scores. The product is available via a web app on a freemium model and supports six languages at launch, including English, Mandarin Chinese, French, Japanese, Korean, and Spanish.

The two most distinctive features are its inline diff review system and an extended thinking mode. The diff review presents every AI-generated edit as tracked changes — additions and deletions displayed side by side — that users can accept or reject with a single click, keeping humans in control of the final document. The thinking mode surfaces the AI's <a href="/news/2026-03-14-multimind-ai-local-first-multi-llm-debate-synthesis">chain-of-thought reasoning</a> in real time before delivering a final answer, a capability particularly useful for complex research queries. Under the hood, DoXmind uses Yjs, a tree-based CRDT library, as the data structure for real-time collaborative editing — a technically sound choice that enables concurrent multi-user editing without requiring a central coordination server.

The editor also ships CSV data analysis and visualizations, semantic search across the knowledge base, structured database views (table, board, gallery, and list), collaborative inline comments anchored to specific text passages, and a presentation mode that converts any document into a full-screen slideshow. Export options include Markdown, PDF, and Word. Aixs Inc. appears to have been founded circa 2024, based on copyright dating on the product website, and its go-to-market strategy — launching via Hacker News's "Show HN" format — signals a technically oriented founding team targeting developer and knowledge-worker communities. It goes up against Notion AI, Coda AI, and Mem.ai; whether the diff-review UX and depth of AI reasoning are enough to pull users away from those entrenched platforms is the question Aixs now has to answer.