Features
Chat with any LLM (OpenAI, Anthropic, Google, local via LM Studio/Ollama)
Vault QA with cited answers from notes and live sources
Projects mode: isolated workspace per project with own model and prompt
Composer: rewrite/expand notes with diff review before accepting
Inline Quick Ask command at cursor, right-click menu, command palette
Context: notes, folders, highlighted text, PDFs, images, URLs, YouTube transcripts, tweets
AI agent with web search, document reading, and long-term memory (Plus)
Semantic and lexical vault search with embedding model support
Custom prompts and commands as Markdown files
Self-host mode using local models for full privacy
Multi-model support: switch providers without lock-in
Relevance: surface related notes you forgot you had
Agent mode v4: autonomous multi-step reasoning and tool use
PDF, EPUB, and 50+ file type support in projects
Everything stored as plain Markdown files – no lock-in
Ask questions across uploaded PDFs and documents
Answers with highlighted source citations
Unlimited questions per document
Rewrite summaries until satisfied
Embed PDF AI in any webpage with a single click
256-bit SHA encryption at rest
Role-based access control
OCR for images and scanned text (Team+ plans)
Response personalization (Team+ plans)
Top performer on RepLiQA benchmark
Department & folder level permissions (Team+ plans)
SOC-2 certificate (Enterprise plan)
