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
Chat with PDF, Word, PowerPoint, Markdown, TXT files
Cited sources with direct page references
Multi-file chats in one conversation
Side-by-side view of chat and PDF
Summarize academic papers, reports, and books
Translate PDFs to any language
YouTube chat (paste link to video)
AI Writer for drafts and outlines
AI Detector for AI-generated content
Flashcard generation from PDFs
Slide generation from PDFs
Research folder organization
Dynamic routing between GPT-4o and GPT-4o-mini
Desktop app and mobile app available
