Explore upcoming capabilities, format support, and engineering milestones for Mai Read.
Every feature is designed with a strict zero-cloud, 100% client-side privacy architecture — your documents never touch a remote server.
Current Releasev1.0 (Live & Stable)
Next Target Milestonev1.1 (EPUB • Dec 2026)
Architecture100% Client-Side / PWA
Community DrivenVote on Priorities Below
In Active Development
Expected: December 2026v1.1
Global Multilingual Experience & International SEO (14+ Languages)#
Read, annotate, and explore in your native tongue. Features comprehensive client-side localization across 14+ major world languages
(French, Spanish, Hindi, Tamil, German, Portuguese, Italian, Japanese, Simplified Chinese, Russian, Arabic, Bengali, and Indonesian)
with automatic browser locale detection, bidirectional Right-to-Left (RTL) typography, native script settings selector,
and search-engine-ready hreflang discoverability.
Pure client-side zero-latency translation catalogs with 100% offline PWA precaching
Automatic browser location/locale detection via navigator.languages with fallback hierarchy
Google Search Central compliant bidirectional hreflang tags & multilingual XML sitemaps
Native script language switcher in Settings and Nav, plus Arabic RTL mirrored ergonomics
Read electronic books (.epub and .mobi) directly in your browser with zero file conversion and zero cloud uploads.
Provides responsive, reflowable typography, chapter-by-chapter table of contents navigation, custom reading margins,
and multi-color text highlighting that persists across browser sessions in local IndexedDB.
Pure client-side WebAssembly & JS rendering with zero server uploads
Reflowable text with serif, sans, and dyslexic-friendly font scaling
Persistent annotations linked directly to study notebook binders
Immediate offline caching for flights, transit, and rural study
Implementation Progress85% Completed
Prototype & Benchmarking
Expected: January 2027v1.2
100% On-Device AI Study Companion (WebLLM & WebGPU)#
Generate chapter summaries, test comprehension with instant active-recall quizzes, and clarify difficult concepts
using browser-native WebGPU (Chrome Built-in Prompt API and WebLLM).
Because the model runs entirely on your local graphics hardware, your private notes and confidential PDFs never leave your machine.
Absolute privacy: zero API keys, zero cloud transmission, zero data logging
Instant 1-click chapter summary into study notebook scratchpad
Active recall question generator grounded directly in your highlighted text
WebGPU acceleration with automatic CPU WebAssembly fallback
Make scanned, image-only PDFs and historical public domain books selectable, searchable, and highlightable.
Using multi-threaded WebAssembly (Tesseract WASM), the OCR engine recognizes text on the fly inside browser memory
without sending sensitive documents to third-party recognition APIs.
Automatic detection of non-searchable image-only document pages
Precise text layer projection matching original typography coordinates
Full offline multi-color highlighting support on recognized scans
Background Web Worker execution to prevent interface stuttering
Keep your study notebooks, reading positions, and highlight collections synchronized across your laptop, tablet,
and smartphone using your own private WebDAV server or Nextcloud instance.
All records are client-side encrypted (AES-GCM 256-bit) prior to transmission, ensuring zero vendor lock-in and zero third-party snooping.
Standard WebDAV protocol support (Nextcloud, ownCloud, Koofr, Fastmail)
End-to-End Encryption (E2EE) with passkey derived in local browser
Differential delta syncing to minimize bandwidth usage
Zero proprietary account requirements: your data remains on your server
Implementation Progress35% Completed
In Specification & UX Design
Expected: April 2027v1.5
Spaced Repetition Flashcards & One-Click Anki (.apkg) Exporter#
Transform your highlighted textbook quotes, paper theorems, and study notebook notes into active recall flashcards.
Review them directly within Mai Read using a built-in SuperMemo SM-2 spaced repetition scheduler, or export
ready-to-import Anki deck packages (.apkg) with automatic source book and page citations.
One-click conversion of text highlights into Q&A or cloze deletion cards
Native .apkg package export with embedded media and formula cards
Integrated in-app review session mode with daily study streaks
Direct deep-link from flashcard back to exact document paragraph
Listen to long public domain books, scientific papers, and study guides hands-free during commutes or workout sessions.
Features smooth word-by-word sentence tracking, adjustable playback speeds (0.75x to 3.0x), offline voice synthesis,
and automatic page advancement that preserves your exact reading position.
Synchronized real-time reading cursor following audio playback
Browser Web Speech Synthesis API with offline WASM neural voice fallback
Background playback support on mobile PWA with lock-screen media controls
Custom speed stepping (1.25x, 1.5x, 2.0x, 2.5x) for speed listeners
Compare two research papers side-by-side or read an academic manuscript alongside its figures, mathematical proofs,
and bibliography simultaneously. Designed specifically for wide desktop monitors, laptops, and horizontal tablet reading surfaces.
Independent zoom, pan, and scroll controls for each split pane
Optional synchronized scroll lock for translated texts and source code
Fast cross-document quote dragging into active study notebooks
Zero canvas memory leak guarantees with dynamic page discarding
Interlink your documents, highlighted passages, and study notebook entries using [[wiki-style]] bracket notation.
Explore an interactive 2D physics-based knowledge graph that visualizes conceptual relationships, common themes,
and cross-disciplinary connections across your entire private offline library.
Bi-directional backlink discovery across books and study notebooks
Interactive HTML5 Canvas force-directed graph with physics simulation
Color-coded nodes by category, notebook, and era
100% client-side indexing with zero server dependencies
Implementation Progress15% Completed
Recently Shipped Milestonesv1.0 Live • October 2026
Organize complex research projects by binding multiple PDFs, DOCX files, and book chapters into unified notebooks.
Features an instantaneous top document switching strip and an auto-saving markdown scratchpad with academic citations (APA, MLA, Chicago).
Read, review, and print Microsoft Word documents without installing Microsoft Office or paid desktop suites.
Decodes Word XML structures natively into beautiful typography right in your browser memory.
Federated catalog discovery across 100,000+ public domain works from Project Gutenberg, the Internet Archive,
academic arXiv preprints, and classic cultural archives with zero account registration.
Permanently bake your multi-color text highlights and drawn margin boxes directly into your PDF file using
pure client-side PDF-Lib binary synthesis. Works offline with zero server processing.
Have a Format or Tool in Mind?
Mai Read is built for researchers, book lovers, students, and autonomous thinkers.
If there is a document format, study workflow, or privacy-first integration you'd love to see on this roadmap, tell us.
Native EPUB 3 and MOBI support is in active development with an expected release of December 2026 in release v1.1.
It runs entirely inside your web browser with zero server uploads, offering responsive text reflow, custom margins,
font choices, and persistent highlighting.
No. In strict accordance with Mai Read's privacy architecture, our upcoming AI Study Companion executes 100% locally on your computer via WebGPU (Chrome Built-in Prompt API and WebLLM). Your notes, reading history, and document contents are never transmitted over the internet or logged on any external servers.
Yes. The client-side OCR engine scheduled for February 2027 in v1.3 uses WebAssembly (Tesseract WASM) running in Web Workers. It turns historical scans and image-only PDFs into selectable, searchable text locally without requiring an internet connection.
We prioritize upcoming releases based on direct community votes on this page, research utility for students and academics, and technical feasibility under our zero-cloud, 100% local privacy guarantee.
No. All core features on our public roadmap remain free, open access, and ad-free. No accounts, credit cards, or logins will ever be required to read documents or use study notebook binders.
Settings & Storage
Preferences & Privacy
Appearance
Content & Privacy
Safe SearchFilter explicit materials in public domain library