v1.8.0 — Recent
- 🎬 Bilibili subtitle import: Extract Bilibili video subtitles. Supports creator spaces, favorites, series/seasons
- 🧩 Side panel redesign: Popup upgraded to Side Panel — opens in place, auto-detects current site and matches import tool
- 🔄 Auto site detection: Auto-match Bilibili/YouTube/Podcast/AI Chat/Web — no manual tab switching needed
- 👤 Google account switching: Switch NotebookLM Google accounts from the side panel header
- 🎓 Onboarding tour: 5-step guided tour after installation, covers all core features
- 📋 Settings panel: New settings page with auto-rename toggle, replay tour, and more
- 📊 Import history: View all import records with easy retry
v1.7.5 — 2026-04-23
- 🎨 Share card preserves AI conversation structure: The previous extractor flattened paragraphs, lists, blockquotes, and bold formatting into a single run-on paragraph — multi-line answers (translations, side-by-side comparisons, definition lists) became unreadable in the share image. Claude/ChatGPT/Gemini extractors now emit cleaned HTML fragments which the offscreen Turndown batch-converts to Markdown, preserving paragraphs / headings / ordered & unordered lists / blockquotes / bold & italic / inline code / fenced code
- 🛡️ Share card renders via marked AST: Walks markdown tokens into React nodes directly, avoiding raw HTML injection paths. Stray iframe/script tags that survive Turndown degrade to plain text automatically
- ✨ NotebookLM imports improve in step: Conversation content imported into NotebookLM now retains the original Markdown structure, much nicer to read
- 🔧 Popup selection preview strips Markdown symbols so
**and##no longer leak into the list
v1.7.4 — 2026-04-09
- 🐛 Fix broken auto-rename for pasted sources: NotebookLM recently started sorting the source list alphabetically, which broke our old assumption that “new source == last in DOM order”.
Pasted Textdefaults were never being renamed to the real title. Switched to a WeakSet diff to locate the new source, robust to any sort order. - 🏷️ Emoji prefixes for rescue / repair: Rescued sources now display as
🛟 Article Title; WeChat/X fake-success sources that were repaired display as🔧 Article Title— so you can tell at a glance which sources have been processed. - ⚙️ New “Auto-rename default-named sources” toggle: Popup → More panel → Settings card. Turns the entire auto-rename behavior on/off dynamically.
v1.7.0 — 2026-03-24
- ▶️ YouTube Import: Import individual videos, playlists, and channel videos into NotebookLM
- Auto-detects current YouTube tab URL for one-click import
- Selective batch import for playlists/channels, with option to limit to the most recent N videos
- Dual extraction strategy: InnerTube API + RSS
- 🗑️ Bulk Delete Revamp: Uses NotebookLM’s native checkboxes, with a new “More Actions” entry at the bottom of the source list
- Confirmation dialog before deletion to prevent accidental removal
- Supports SPA in-page navigation (no page refresh needed)
- 🎓 Onboarding guide now includes YouTube step
- 🏠 Welcome page adds YouTube and More feature cards
v1.6.8 — 2026-03-22
- 🔧 Fix repair/rescue tab targeting issues
- 📊 Switch to port-based progress communication
- 🛡️ Banner state protection: in-progress banners no longer get removed by DOM mutations
v1.6.7 — 2026-03-21
- 📌 Add “Pin to toolbar” guidance tooltip on Welcome page
v1.6.6 — 2026-03-20
- 🎯 Notebook target selector: choose a notebook from the dropdown and imports go directly there — no manual tab switching
- 🚀 One-click import button: prominent yellow Import button on the bookmarks page for importing the current page to NotebookLM
- 🎓 Onboarding: Welcome page + 5-step tooltip guide on install, covering all 4 core features, skippable and replayable from the More page
- 🎨 Bookmarks panel color refresh: bookmark-related UI unified to blue, Import button highlighted in yellow
- 🔧 Fix import opening homepage: initially auto-selected notebook now correctly persists to storage
- ⚡ Notebook list caching: 5-minute TTL cache, skeleton loader on load, manual refresh to force update
- 📐 Fix popup viewport overflow: add max-height constraint to prevent content overflow
v1.5.8 — 2026-03-11
- 🎴 AI conversation share cards: export conversations as beautifully designed image cards with warm book-style typography
- Default export as JPEG; dropdown offers PNG, PDF, and clipboard copy
- iPhone Dynamic Island safe area support (59pt, per Apple HIG)
- QR code + bilingual signature (Made with ❤️ by Flow2Note)
- 3x Retina high-resolution rendering
- 🖨️ PDF export style upgrade: warm book-style design (Noto Serif SC + EB Garamond), cream background + russet accent
- Footer signature area: brand logo + YouTube link + bilingual signature
- System font fallback chain for offline rendering (Georgia → Songti SC → STSong → SimSun)
- 📋 Clipboard copy button added to Bookmarks and Doc Sites: small icon button next to PDF export, one-click copy Markdown to clipboard
- 📝 New GitHub Issue templates: Bug Report / Feature Request / Site Support Request
v1.5.7 — 2026-03-09
- 🖨️ Fix X.com article PDF export: correctly extract article body, support CJK Chinese fonts, fix paragraph merging issues
- 🎨 PDF reading experience upgrade: warm book-style typography (Georgia serif font, cream background, russet accents)
- 📄 Single article PDF optimization: skip table of contents and cover pages, use article title as filename
- 📝 Auto-insert heading in PDF content: automatically prepend h1 heading to each article (skip if heading already exists)
- 🔧 X.com extraction improvements: Article → tweetText dual-strategy fallback, render wait time 5s → 8s
v1.5.6 — 2026-03-08
- 🗑️ Bulk source deletion:
Cmd+Shift+S(Mac) /Ctrl+Shift+S(Windows) enters selection mode — supports select all, invert selection, select failed only, and one-click bulk delete - 🛟 Improved error detection: MutationObserver watches for class attribute changes to detect failed sources in real time (#35)
- 🔄 Delayed re-check: auto re-checks at 5s/10s/18s after import to catch “false success” sources like WeChat articles
- ✏️ Fix smart rename for pasted text sources (#38)
- 🏗️ Huawei developer docs support: fix Catalog API parsing + tab-based content extraction (#37)
v1.5.5 — 2026-03-02
- 🌐 Complete i18n for content script banners and popup components
v1.5.4 — 2026-03-01
- 🌐 Add Chrome i18n support with
_localesdirectory (en / zh_CN)
v1.5.3
- 📝 README update: screenshots, YouTube demo video, English content
v1.5.2 — 2026-02-25
- 🌐 Add X.com (Twitter) long-form article import via tab-based rendering for full content extraction
- 🛟 Rescue feature expanded: supports failure detection and one-click repair for X.com dynamic pages
- ⚙️ Refactor tab-based extraction pipeline, unifying WeChat / X.com and other SPA page handling
v1.1.43 — 2025-02-24
- 🧹 Remove dead YouTube-related code
- 🎨 Add store promotional graphics
- 📊 Tests: all 84 tests passing
v1.1.42 — 2025-02-24
- 🌐 Complete i18n for all components (ImportPanel was the last one)
- 🔄 Add manual language toggle (EN/中) with localStorage persistence
- 🤖 AI Conversation tab: add 4-step onboarding guide
- 🎨 Update extension icon: blue circle + white jetpack figure
- 🧪 Add 37 tests (bookmarks 21 + i18n 11 + podcast 6), bringing total to 99
- 📄 Add store listing docs and privacy policy
v1.1.41 — 2025-02-24
- 🌐 Create i18n system (
lib/i18n.ts) witht()translation function - 🔄 11 components converted to i18n
- 🌍 Auto-detect
navigator.language
v1.1.40 — 2025-02-24
- 📚 Verify English NotebookLM DOM matching
- 🔧 Fix text matching for “Websites” (plural), “Paste any links” (plural), etc.
v1.1.39 — 2025-02-24
- 📑 Move Bookmarks tab to first position (default tab)
- 📖 Add Bookmarks onboarding guide
- 📂 Separate dev/release build directories:
dist-dev/vsdist/
v1.1.38 — 2025-02-24
- 🔍 Language-agnostic button matching: based on Material Icons instead of text
- 🎯
findDialogButtonByIcon()supportslink,content_paste,arrow_back,add,edit - 📝 Text matching retained as fallback
v1.1.37 — 2025-02-24
- ⏱️ Improve rename reliability: 4s initial wait, 6 retries, incremental delay (600ms → 2600ms)
- 🌍 Multilingual default source name matching