One memory, every feature
The same per-domain memory powers the tutor, document chat, and cross-library analysis — not a fresh blank context each time.
Notedeck ingests your books, papers, and notes — then teaches you from them, quizzes you on them, and shows you the structure between them.
See it in action
Real product footage: knowledge graph, question bank, cross-bank analysis, and the AI classroom.
Real-time slides, narrated lessons, raise-your-hand interruptions — all generated from what you've uploaded. Not a curriculum. Yours.
Once you have more than one knowledge base, Notedeck holds them side by side — finding bridges between fields, contradictions between sources, ideas one library is hiding from the other.
Under everything sits a holistic memory — one per knowledge domain, shared by the AI tutor, your document chats, and analysis. It extracts what landed and what tripped you up from each interaction, recalls what is relevant, and tunes explanation depth to you. Long-term, cross-channel, and personal — the way Claude Code remembers a codebase, Notedeck remembers how you learn. Generic AI chat forgets you every session; this is vibe learning.
The same per-domain memory powers the tutor, document chat, and cross-library analysis — not a fresh blank context each time.
Every session is mined for what you grasped and where you stumbled, then recalled by relevance when it matters.
It reads your signals and meets you where you are, so the next lesson starts where you actually left off.
After chatting across the tutor and your documents, the answers scatter into endless scroll-back. The Study Wiki is the human-readable face of your memory: an LLM keeps weaving your Q&A into structured, interlinked knowledge pages — grouped by domain (physics, finance, work…), built only from what you actually asked, never scraped from the source files. The systematic record a chat with ChatGPT or Claude never leaves behind.
Compiled from your Q&A — not your source files
Built from what you asked, never scraped from the uploaded files — it mirrors your thinking, not the source.
Questions scattered across sessions become coherent, interlinked pages, one knowledge domain at a time.
It is the browsable face of the same memory that drives the tutor — one click keeps it fresh.
Notedeck is an AI workbench that turns books, papers, and notes you bring in into a knowledge graph you can chat with, learn from, and quiz yourself on.
ChatGPT answers from its training data. Notedeck answers from YOUR materials — your textbooks, papers, lecture notes — and shows you where each answer came from. It also builds a knowledge graph across all your documents so you can see how concepts connect.
Notedeck adds an AI tutor that quizzes you, a knowledge graph that spans all your uploaded documents (not just one notebook), and explicit spaced-repetition learning workflows. NotebookLM is a great Q&A tool over a single notebook; Notedeck is a workbench for a whole personal library.
PDF, DOCX, PPTX, EPUB, plain text, and Markdown. Images and scanned PDFs are processed with vision models for text and diagrams.
No. Your materials are used only to answer your own questions and build your own knowledge graph. We never use them to train base models.
Yes. The Free tier covers your first 1 GB of storage, ~ 100 documents, and unmetered chat and learning sessions on those documents.
The product UI ships in English and Chinese today, with translations for Japanese, Korean, French, German, Spanish, Brazilian Portuguese, Turkish, Italian, and Arabic rolling out. AI answers and document understanding work in any major language regardless of UI language.
Yes. Documents are downloadable in their original format at any time. Notes, annotations, and quiz history export as JSON or Markdown.
Free tier, Pro tier (USD 20/month or 192/year) for higher storage and advanced features, and Max tier (USD 50/month or 480/year) for power users. See the Pricing page for full details.
United States. Cross-border transfer disclosures and data processing details are in our Privacy Policy.