🚀 What's new
Intelligence that
lives where you do.
VaultBook 3.0 brings machine learning, neural search, and adaptive AI — all running locally in your browser. No servers. No latency. No compromise. Your workspace now thinks with you.
VaultBook 3.0 — Released February 2026
🧠 AI & Machine Learning
Your notes now learn from you
Every interaction teaches VaultBook what matters to you — and it remembers, adapts, and surfaces the right thing at the right time.
✨
AI Suggestions Pager
A 4-page adaptive carousel that learns your behavior. Page one predicts what you'll need next based on upcoming scheduled entries and weekday reading patterns across the last 4 weeks. Pages two through four surface recently read entries, recently opened files, and recently used tools — a personalized command center built from your own neural pathways through your library.
Adaptive ML
🔎
Neural QA Search
Ask your library a question in plain English. VaultBook's weighted scoring engine searches across 7 content layers — titles (weight 8), labels (6), inline OCR text (5), body content (4), sections (3), main attachments (2), and section attachments (1) — then ranks results through a multi-field scoring pipeline that gets smarter with every vote.
Deep NLP
🔗
Related Entries with Similarity Learning
When you open any note, VaultBook's contextual similarity engine finds entries you didn't know were connected. Reddit-style upvote/downvote on each pair trains a persistent relevance model — your votes reshape the neural map of your knowledge base over time. The system learns the associations that matter to you, not just keyword overlaps.
Reinforcement Learning
📊
Vote-Based Reranking
Every QA search result can be upvoted (+1M) or downvoted (-1M), persistently stored in the userVotes schema. Over sessions, your feedback creates a personalized relevance layer that sits on top of the scoring engine — turning VaultBook from a search tool into a system that understands how you think.
Persistent Learning
⚡ Deep Document Intelligence
Every byte is searchable
VaultBook doesn't just store files — it reads them, extracts their content, and weaves them into your searchable knowledge graph.
📄
PDF Intelligence
Text layer extraction via pdf.js, plus OCR of rendered pages for scanned documents. Every PDF becomes searchable instantly.
📊
XLSX & PPTX Parsing
SheetJS extracts spreadsheet cell data, JSZip reads slide text from presentations. Structured content feeds directly into the search index.
📦
Archive Crawling
ZIP files are opened, their text-like inner files indexed, and embedded images run through OCR — all without extracting to disk.
📧
Outlook MSG Parsing
Subject, sender, body text, and nested attachments are extracted and deep-indexed. Your archived emails become part of your knowledge base.
👁️
Inline OCR Engine
Images embedded in entries are automatically OCR'd, cached per-item, and indexed. The top 12 QA results trigger background OCR warm-up for speed.
🔬
Cross-Format OCR
OCR reaches inside DOCX (word/media), XLSX (xl/media), and ZIP archives to extract text from images embedded deep inside compound documents.
🛠️ Built-in Tools
14+ tools. Zero context-switching.
Everything you'd normally open a separate app for — built directly into your workspace. Process, analyze, and create without ever leaving VaultBook.
📋
Kanban Board
Auto-creates columns from your labels and inline #hashtags. Drag-and-drop cards that sync back to your notes in real time.
📰
RSS / Atom Reader
Subscribe to feeds, organize into folders, and pull articles directly into your workspace without leaving VaultBook.
📑
PDF Merge, Split & Compress
Combine multiple PDFs, extract page ranges, or compress scanned documents — all processed locally in the browser.
🎵
MP3 Cutter & Joiner
Trim silence, cut clips, and join audio segments. Perfect for processing voice memos, interviews, and podcast clips.
📁
File & Photo Explorer
Browse attachments by type, entry, or page. Scan entire folders of photos and videos with a visual grid interface.
📈
File & Folder Analyzer
Analyze CSV and TXT data files with instant visualizations. Audit disk space and file sizes across your storage folder.
🔑
Password Generator
Generate cryptographically strong passwords with one click. Copy instantly, no external password tool needed.
🔗
Save URL → Entry
Paste any URL and VaultBook creates a structured note from the web page content — instant web clipping.
💬
Threads
Chat-style notes in a centered overlay. Perfect for brainstorming sessions, quick captures, and conversational note-taking.
🔮 The road ahead
Where VaultBook is going
Every release pushes the boundary of what a local-first, browser-based workspace can do. Here's what we're building toward.
🧬
Semantic Neural Index
Moving beyond keyword matching to true semantic understanding — connecting ideas by meaning, not just shared words, all computed locally on your device.
🌊
Adaptive Workflow Engine
VaultBook will learn your workflows — which tools you open after which searches, which pages you visit in sequence — and begin anticipating your next move.
⚡
Edge ML Acceleration
Leveraging WebGPU and WebAssembly to run larger models directly in the browser — faster OCR, richer similarity scoring, and real-time document understanding.
📋 Changelog
Release history
A record of every major milestone from 2020 to now.
Feb 2026
VaultBook 3.0
AI Suggestions pager with personalized 4-page carousel. Timetable and calendar views with day/week layouts. Canvas-rendered analytics dashboards with label, page, and activity charts. 60-day version history. Refined UI with attachment ticker and random note spotlight.
2025
Deep Indexing & Pro Tools
Deep file indexing for XLSX, PPTX, PDF, ZIP, and MSG. Cross-format OCR inside DOCX, XLSX, and archive images. 14+ built-in tools including Kanban, PDF suite, MP3 editor, RSS reader, and file explorers. Multi-tab views with advanced filters.
2024
Vote-Based Learning & Related Entries
Related entries with contextual similarity suggestions. Reddit-style persistent vote system for both search results and related entries. Smart label suggestions with content analysis. Attachment type bucketing and file extension grouping.
2023
Team Scale-Up & Rich Editor
Team grew to 50 specialists. Rich text editor with tables, code blocks, callout blocks, H1–H6 headings, font selectors, and case transforms. Sections with sub-entries, each with their own attachments and rich editor. Hierarchical pages with nested notebooks.
2022
AI Search Goes Live
Weighted search scoring across 7 content layers. Natural-language QA mode with paginated results. Typeahead search with real-time dropdown. Inline OCR for automatic image text extraction. Query suggestions from search history.
2021
Encryption & Storage Layer
AES-256-GCM per-entry encryption with PBKDF2 key derivation at 100,000 iterations. File System Access API integration with repository.json and sidecar .md files. Lock screen with full-page blur overlay. Autosave with dirty-flag debouncing.
2020
The First Commit
VaultBook begins as a single HTML file — a personal tool to organize research notes offline. Local-first storage, zero dependencies, and the principle that your data belongs to you.
The smartest workspace that never phones home.
All this intelligence. All running locally. All yours.