更新日志

记录 Web2MD 的每一次进步

v1.1.2May 9, 2026Latest

NewBulk-export AI conversations to a single Markdown file. The new queue workflow lets you collect multiple conversations across ChatGPT, Claude, DeepSeek, Gemini, and Perplexity, then download them all in one merged .md — perfect for migrating from one AI to another, or archiving research into Obsidian/Notion. Works the same on every platform; no broken list APIs

NewMode switcher on AI platforms — when you're on chatgpt.com, claude.ai, etc., the popup shows "Send page → AI" vs "Export $platform$ chats" as two visible modes. Both workflows are explicit instead of guessing intent from URL

NewQueue persists across popup close — switching browser tabs to grab the next conversation no longer loses what you've collected. Items survive popup open/close cycles via chrome.storage.local

ImprovedVisual reinforcement on copy — the "Copied to clipboard" toast now uses accent color, two lines, and 50% longer display. First line confirms the action ("Copied — ready to paste"); second tells you the next step ("Press ⌘V or Ctrl+V in any AI chat to paste"). Surfaces the cross-platform paste workflow that was previously hidden

ImprovedBetter conversation titles in the queue — previously every queued item showed as "Untitled" because AI chat extractors put the conversation title in the markdown body, not metadata. Five-tier title fallback chain (metadata → first H1 → first text line → URL slug → "Untitled") now derives meaningful titles automatically

FixedPro user activity tracking — four conversion paths (Pro popup saves via /api/history, /api/v1/convert, /api/v1/batch, /api/convert-pdf) were silently skipping the UsageRecord write, making Pro users invisible to the retention dashboards and email triggers. All four now route through one usage-tracking helper. Closes the analytics blind spot that was suppressing renewal emails for paying users

FixedAPI daily-limit gate — Free users with API keys could call /api/v1/convert and /api/v1/batch unlimited times because the limit check was missing. Both endpoints now enforce the per-plan daily cap and return 429 with a useful error when exceeded

FixedWebhook reliability — the Creem checkout webhook was using fire-and-forget for post-checkout work (welcome email, getting-started email, funnel events). Vercel freezes the function as soon as it returns, killing in-flight promises. Switched to awaited try/catch so all three side-effects complete before the webhook returns 200

v1.1.0May 7, 2026

ImprovedCode block, table, and LaTeX fidelity — DOM preprocessing strips Highlight.js / Prism / Shiki residue spans, unwraps GitHub gist-style line-number tables, expands colspan, and converts KaTeX / MathJax renders back to TeX source. Wikipedia, arXiv, and Stack Overflow conversions now preserve formatting cleanly in Obsidian and Claude — fixes the #1 complaint about every other Markdown clipper

NewAI conversation export — dedicated extractor for ChatGPT (chatgpt.com/c/), Gemini (gemini.google.com), DeepSeek (chat.deepseek.com), and Perplexity (perplexity.ai/search) conversation pages. Exports the full multi-turn dialog as Markdown with ## User / ## Assistant headings. Joins the existing Claude chat extractor

NewComparison pages for MarkDownload, Pocket, Firecrawl, and Obsidian Web Clipper — research-backed feature comparisons at /alternatives/[name]. Helps users coming from those tools see what Web2MD does differently

FixedPro user onboarding — webhook now sends an extension-signin nudge email alongside the welcome receipt. Dashboard shows a clear banner when a Pro user has 0 conversions, with step-by-step instructions to sign in inside the extension. Fixes the silent activation gap where paid users couldn't see their data because the extension was using anonymous IDs

New6 new blog posts targeting high-intent search terms surfaced in our competitor research: Xiaohongshu-to-Markdown, AI conversation export, $9 Firecrawl alternative, why ChatGPT copy-paste looks terrible, web clipper landscape after MarkDownload + Pocket exit, and WeChat 公众号 to Markdown

v1.0.6May 1, 2026

NewXiaohongshu (小红书) extractor — pulls note title, body, author, IP location, tags, images, and engagement stats. AI search engines (ChatGPT, Perplexity, Gemini) cannot scrape Xiaohongshu; Web2MD reads the page from inside your browser, so it works where every server-side scraper fails

NewWeChat Official Account (微信公众号) extractor — extracts the canonical article body with lazy-image hydration. Tencent blocks every external fetch, so this is the cleanest way to get WeChat content into AI tools

NewZhihu (知乎) extractor — supports column articles (zhuanlan.zhihu.com/p/), single answers, and full questions (top 5 answers sorted by votes)

NewBilibili (B站) extractor — for videos: title, UP主, description, tags, view/like/coin/danmaku stats. For column articles (cv): full body extraction

NewSite prompt banners — when you land on Xiaohongshu, WeChat, Zhihu, or Bilibili, a slim hint shows the keyboard shortcut. Auto-dismisses in 5 seconds, max 3 times per site

v1.0.5April 30, 2026

FixedSend to AI on Claude restored — recent ProseMirror changes broke the injection. New multi-strategy injector (paste event → execCommand → DOM fallback) verifies the text actually landed before reporting success, so silent failures and "fallback to clipboard" cases drop sharply

FixedArc browser support — the extension icon was effectively dead on Arc because it exposes the chrome.sidePanel API but has no side panel UI. Default UI mode is now Popup for new users; existing Chrome/Edge sidebar users are unaffected

NewHelpful prompts on AI platforms (Claude, ChatGPT, Gemini, Perplexity, Grok, DeepSeek) — a small reminder shows when you land on these sites so you remember Web2MD can pull a webpage in. Auto-dismisses in 5 seconds, max 3 times per site, then 7-day cooldown

NewSmart extractor hints on Reddit, X/Twitter, GitHub, YouTube, Hacker News, arXiv, Medium, Substack, Wikipedia, Dev.to, Product Hunt, Stack Overflow, Notion, Feishu, and OpenAI Docs — when you visit a site Web2MD has a dedicated extractor for, you get a one-line nudge with the keyboard shortcut

NewQuick in-product survey — after your first 3 conversions, Web2MD asks how you mainly use it. One question, one tap, never asked again. Helps us ship features that match how you actually work

ImprovedSide Panel toggle in Settings now indicates when the API is unsupported (Arc, etc.) and routes to Popup automatically — no more silently broken icon click

v1.0.4April 19, 2026

NewSide Panel mode — Web2MD now opens as a persistent sidebar alongside the webpage. Stay open while browsing, no more popup closing on click. Toggle between Side Panel and Popup in Settings

NewFirefox Add-ons support — Web2MD is now available on Firefox (v128+) with full feature parity except Side Panel

NewDownload is now a Pro feature — download .md files requires Pro plan, the #1 upgrade trigger based on user data

ImprovedLocalized Chrome Web Store descriptions — full native-language store listings for Japanese, Korean, German, French, Spanish, Portuguese, Russian, Arabic, and Traditional Chinese

Improved19 missing i18n keys added — addToQueue, clearQueue, sendAllToAI, tooltips, and other UI text now properly translated across all 11 languages

v1.0.3April 18, 2026

NewNotebookLM integration — send converted Markdown directly to Google NotebookLM from the preview card, result panel, and batch convert panel

NewFree users now get 5 saved history items — view your recent conversions in the dashboard without upgrading

ImprovedConversion success rate improved — 15 new content selectors added (Wikipedia, Stack Overflow, Dev.to, news sites), smarter fallback extraction reduces failure rate from 18% to under 4%

ImprovedUnified free tier — 3 conversions per day for all users, simpler and more predictable

FixedFixed payment flow — clicking Upgrade while not logged in now correctly redirects back to pricing after login and auto-triggers checkout

FixedReduced extension uninstall rate — softer paywall messaging, delayed login prompts for new users

v1.0.1April 12, 2026

NewAgent Bridge — AI agents (Claude Code, Cursor, Cowork) can now control the Chrome extension via Native Messaging to batch-convert Reddit, JS-rendered, and login-protected pages without manual interaction (PRO)

NewAgent Bridge setup in Settings — step-by-step instructions for installing the native messaging host and configuring MCP in Claude Code / Cursor

NewMCP tools: agent_convert (single URL) and agent_batch_convert (up to 50 URLs) — AI agents call these tools and the extension opens tabs, extracts content, and returns Markdown automatically

ImprovedSite-specific extraction now runs before opening a tab in batch conversions — Reddit, Hacker News, and other API-based extractors skip the tab entirely, reducing latency and avoiding timeouts

ImprovedSettings page now shows real user avatar instead of initial letter

FixedFixed native messaging host not launching on macOS due to TCC restrictions on ~/Desktop — host now installs to ~/.web2md/

v0.9.9April 8, 2026

NewBatch API — convert up to 50 URLs in a single request via POST /api/v1/batch. Returns structured results with per-URL success/error and token counts (PRO)

NewHosted MCP — zero-install Claude Desktop integration via /api/mcp. Add one URL + API key to Claude Desktop config and Web2MD becomes an AI tool instantly (PRO)

NewToken compression stats — dashboard now shows total tokens processed and average compression ratio, quantifying how much HTML Web2MD strips per conversion

NewCLI compression display — npx web2md now shows word count, token count, and compression ratio in every Done message (e.g. "3,821 tokens · 38x compression from 145KB")

NewExtension Bridge — CLI flag --bridge routes conversions through your Chrome extension, giving command-line tools access to browser rendering and login state (WEB2MD_EXTENSION_ID=xxx npx web2md URL --bridge)

NewDashboard Settings page — manage API keys (up to 5) and webhooks (up to 3) directly from the dashboard at /dashboard/settings

ImprovedUnified dashboard navigation — consistent header across Knowledge Base, History, and Settings pages with active-state highlighting

FixedFixed API key and webhook creation from the extension settings panel — Bearer token auth was not accepted by the API routes, causing silent failures

v0.9.8April 4, 2026

NewSyntax highlighting in Markdown Editor — 20 languages with One Dark theme (TypeScript, Python, SQL, Rust, Go, and more)

NewKaTeX math rendering — inline ($E=mc^2$) and block ($$...$$) LaTeX formulas render live as you type

NewMermaid diagram support — flowcharts, sequence diagrams, pie charts, and more render directly in the preview

NewTable of Contents — auto-generated from headings with click-to-scroll navigation, shown in full-preview mode

ImprovedMarkdown Editor now at /tools — cleaner URL, easier to remember and share

ImprovedShared rendering engine — web app and Chrome extension now use the same markdown-renderer for consistent output

ImprovedDashboard URL converter — always visible with a more prominent input box

v0.9.3March 27, 2026

NewKnowledge Base Dashboard — redesigned dashboard with Collections overview, AI Tag Cloud, Source Domains chart, and Recent Knowledge feed with summaries and tags

NewBranded preview card — Copy, Download, and Send to AI buttons now show icon + text in a 2×2 grid layout

NewBranded loading toast — right-click conversion shows .MD logo badge with accent-colored spinner

New14 smart extractors — added Medium, Substack, Hacker News (Algolia API), arXiv, Wikipedia, DEV.to, and Product Hunt

NewTag-based collections — AI-generated tags now shown as a browseable "By Tag" view in Collections tab

NewMulti-select in History — select multiple items and batch-add them to a collection

NewDownload button — save as .md file from History, Convert result, and Preview Card

NewDomain usage tracking — every conversion logs the source domain for data-driven extractor prioritization

ImprovedTwitter/X extractor — only extracts from primary column, filters sidebar and recommendations

ImprovedConversion tooltips — all action buttons (Copy, Send to AI, Download, Collection, Delete) now show descriptive tooltips on hover

FixedFixed crash when clicking Collections, Settings, or "Add to Collection" — API responses with unexpected shape no longer cause white screen

FixedFixed onboarding appearing for existing users after extension update

FixedFixed auth token not syncing — popup now recovers login state from web2md.org session cookie

FixedFixed "Internal server error" crash when convert API returns error — service worker now guards all 3 convert handlers

v0.9.2March 24, 2026

NewMarkdown Editor now available at /md — shorter URL, easier to remember

NewMD → Natural Language conversion — intelligently strips Markdown syntax into clean prose (lists become comma-separated sentences, headings become labels)

NewMarkdown formatting toolbar — one-click buttons for bold, italic, headings, links, images, code blocks, lists, tables, and more

NewKeyboard shortcuts — Cmd+B (bold), Cmd+I (italic), Cmd+K (link), Cmd+Shift+K (code block), Cmd+Shift+Q (quote)

NewDrag & drop .md files — drop any .md or .txt file directly into the editor to load its content

NewToken estimation — real-time GPT-4/Claude token count displayed in the status bar

NewFullscreen mode — immersive distraction-free editing with Escape to exit

NewAuto-save — editor content persists to localStorage, survives page refresh

ImprovedEnhanced Markdown Beautify — smarter formatting with blank line normalization, list marker cleanup, heading spacing, and horizontal rule standardization

ImprovedFormat conversion section now includes 4 tabs: HTML→MD, MD→HTML, MD→Plain Text, MD→Natural Language

v0.9.1March 24, 2026

NewAI-first Convert tab — AI platform selector and prompt templates front and center, one-click "Convert & Send to AI"

NewMulti-page queue — convert multiple pages, then send all to your AI in one merged request

NewAuto AI tags & summary — every conversion auto-generates semantic tags and a 1-2 sentence summary powered by Doubao AI

NewAuto platform collections — conversions from YouTube, Twitter, Reddit, GitHub, Medium etc. auto-sorted into platform-based collections

NewYouTube subtitle extraction — improved support for auto-generated captions (ASR) with timestamped paragraph formatting

NewFeature usage analytics — self-hosted tracking system to understand which features are actually used

ImprovedRedesigned navigation — clean 4-tab layout (Convert / History / Collections / Settings) with icon + label

ImprovedNew prompt templates — "Tweet" (turn articles into X threads) and "Q&A" (generate 5 Q&As) replace less useful templates

ImprovedHistory tab now available to all users (previously PRO-only)

ImprovedOnboarding guide now properly activates on first install (was built but never wired up)

ImprovedWelcome page now includes sign-in CTA (previously only had "Rate" and "Docs" links)

ImprovedAuto-copy to clipboard enabled by default after conversion

FixedFixed AI summary pipeline — was silently failing due to missing API key configuration

FixedFixed free tier limit display showing 20 instead of 3 conversions/day

FixedFixed Chinese tab labels wrapping/breaking in narrow popup

v0.9.0March 22, 2026

NewWatch Mode — monitor any webpage for changes with hourly, daily, or weekly checks. Auto-convert changed content or trigger webhooks (PRO)

NewAI Research Assistant — input a topic and URLs, get a structured research report with executive summary, key findings, and detailed analysis powered by Claude (PRO)

NewSmart Clip Suggestions — browse the web and get automatic "Save to Collection" prompts when page content matches your existing collections (PRO)

NewPDF to Markdown — convert PDF files directly to clean Markdown, with text extraction and scanned document detection (PRO)

NewNew Watch tab in extension popup — manage watch rules, check pages on demand, see change history

NewResearch panel in Collections — launch AI research directly from your collections workflow

ImprovedExtension now detects PDF pages and shows a dedicated "Convert PDF" button

ImprovedBackground watch polling with browser notifications when monitored pages change

ImprovedCollection tags cached for faster smart clip matching

v0.8.0March 21, 2026

NewPublic REST API — convert any URL or HTML to Markdown via /api/v1/convert with API key authentication (60 req/min)

NewMCP Server — use Web2MD directly from Claude Desktop, Cursor, or Windsurf with convert_url, semantic_search, and get_conversion tools. Install via npx web2md-mcp

NewAPI Key management — create up to 5 API keys, view usage timestamps, manage from extension Settings (PRO)

NewWebhook automation — receive POST notifications when conversions complete, with HMAC-SHA256 signature verification and automatic retry (PRO)

NewDashboard History page — browse all past conversions with search, tag filter, pagination, clickable source URLs, and one-click Markdown copy

ImprovedExtension now requires login — sign in once and your conversions sync across devices

ImprovedAPI key authentication now supported across search and history endpoints for full programmatic access

FixedFixed extension session not syncing with dashboard — logging out on the website now logs out the extension too

v0.7.1March 19, 2026

NewNotion page extractor — open any Notion page in your browser and convert it to clean Markdown with a single click. Supports headings, text, code blocks, lists, toggles, callouts, quotes, images, tables, bookmarks, and embeds

ImprovedDual extraction strategy: primary block-level DOM parsing via data-block-id, with container fallback for pages that use non-standard layouts

v0.7.0March 18, 2026

NewCollections — organize conversions into folders like "React Research" or "Competitor Analysis", with nested sub-collections and one-click merged Markdown export

NewSemantic Search — natural language queries powered by pgvector + OpenAI embeddings; ask "articles about database optimization" and find them across languages

NewRelated Content — after converting a page, see up to 3 semantically similar articles you previously saved

NewCustom Extraction Rules — define per-domain CSS selectors for content, removal targets, and title overrides; manage rules in Settings

NewAdd to Collection — save any history item to a collection directly from the History tab with a folder dropdown

ImprovedNew Collections tab in the extension popup with create, browse, export, and delete functionality

ImprovedEmbedding pipeline runs async on every conversion — builds your searchable knowledge base automatically

v0.6.0March 18, 2026

NewAI Smart Summary — every conversion now auto-generates a 1-2 sentence summary and 3-5 semantic tags via Claude Haiku (Pro)

NewBlurred AI summary preview for Free users — see a taste of AI insights with an upgrade CTA on the preview card

NewSelection Toolbar — select text on any page to get a floating toolbar with Copy as MD, Append, and Send to AI buttons

NewAppend Mode — select multiple text fragments across a page, then combine them into a single Markdown document

NewContext Window Optimizer — choose your target AI model (Claude 200K / GPT-4 128K / Gemini 1M) and long content is auto-compressed to fit

NewDate range filter in History — filter saved conversions by start and end date alongside search and tag filters

ImprovedHistory search now covers title, URL, and AI-generated summary for more relevant results

ImprovedSmart content compression preserves code blocks, headings, and first/last sections while trimming middle content

v0.5.1March 8, 2026

NewFeishu/Lark cloud document extractor — converts Feishu Docx and Wiki pages to Markdown with full block model support (headings, code, tables, images, lists, callouts, and more)

NewDual extraction strategy: primary extraction via window.PageMain block model tree, with DOM fallback for pages where the block model is unavailable

FixedFixed X/Twitter Articles (long-form Notes) producing empty content — added full Draft.js block parser supporting headings, lists, code blocks, images, inline styles, and entity links

FixedFixed server-side conversion crash caused by linkedom NodeList not supporting .forEach() — replaced all instances with Array.from() + for...of across shared converter

v0.5.0March 5, 2026

NewDeep content cleaning — 130+ exact CSS selectors and 160+ fuzzy patterns remove ads, navs, sidebars, cookie banners, and social widgets before conversion

NewRich metadata extraction — author, published date, description, and featured image pulled from Open Graph, Twitter Card, Schema.org JSON-LD, and meta tags with multi-layer fallback

NewYAML frontmatter now includes author, published, and description fields when metadata is available

ImprovedRetry with relaxed cleaning — when content extraction yields < 200 characters, the engine retries without fuzzy pattern matching to recover over-cleaned pages

ImprovedTitle cleaning — site name suffixes (e.g. "Article Title | Site Name") are automatically stripped using OG/Schema.org metadata

ImprovedClutter detection now checks 7 attributes (class, id, data-test, data-testid, data-test-id, data-qa, data-cy) instead of just class and id

v0.4.9February 28, 2026

NewFree online Markdown editor tool page — edit, live preview, beautify, and convert Markdown at web2md.org/tools

NewFormat conversion tools — HTML → Markdown, Markdown → HTML, and Markdown → Plain Text, all running client-side

NewCodeMirror editor with Markdown syntax highlighting and One Dark theme

ImprovedMarkdown beautifier — collapses blank lines, normalizes headings, trims trailing whitespace

NewWord count, character count, and line count stats bar

v0.4.7February 25, 2026

ImprovedFake headings using role=heading or class=hN are now converted to proper Markdown headings

ImprovedCopy buttons, line numbers, and toolbars inside code blocks are stripped before conversion

ImprovedTabbed code examples (Docusaurus, VitePress) are fully expanded — all language variants preserved

ImprovedSPA and JS-rendered pages now wait for content to load before capturing HTML

FixedReddit links pointing to a specific comment now fetch the full thread instead of just that comment

v0.4.6February 24, 2026

ImprovedSend to AI for ChatGPT and Perplexity now uses URL parameter scheme (?q=...) — no script injection, instant and reliable

ImprovedPerplexity URL updated to /search/ path for direct query support

ImprovedContent exceeding 8000 characters automatically falls back to injection scheme

v0.4.5February 23, 2026

NewDashboard URL Converter — paste any URL on mobile to convert to Markdown, no extension required

NewServer-side URL fetching — bypasses CORS limitations, works from any mobile browser

ImprovedDashboard mobile layout — 2-column stats grid, URL field hidden on small screens, username hidden on mobile header

v0.4.4February 22, 2026

NewTask list support — checkboxes preserved as - [x] / - [ ] in converted Markdown

NewLaTeX math support — inline ($...$) and block ($$...$$) formulas preserved from span.math, div.math-display, and MathJax script tags

NewFootnote support — superscript references converted to [^N] and footnote definitions to [^N]: ...

NewMermaid diagram support — div.mermaid blocks wrapped in ```mermaid fenced code blocks

v0.4.3February 22, 2026

NewObsidian export now auto-generates YAML frontmatter (title, source, author, date, description) — notes are immediately queryable in Obsidian

NewMetadata extraction: author, published date, and description pulled from OG tags, meta tags, and JSON-LD — available in frontmatter without extra steps

ImprovedObsidian export switched to clipboard channel — no more 2 MB size limit; falls back to URI transfer if clipboard is unavailable

ImprovedSettings now sync across devices via chrome.storage.sync — vault name, AI platform, and all preferences restored automatically after reinstall or on a new machine

v0.4.2February 21, 2026

ImprovedReplaced Mozilla Readability with a semantic-first content extractor — better results on marketing pages, docs sites, and product pages

ImprovedSingle-pass HTML parsing instead of three — faster extraction with less memory overhead

FixedEliminated class-string matching that incorrectly removed content containing words like "comment" or "sidebar"

v0.4.1February 21, 2026

FixedKeyboard shortcut now works reliably via content script keydown listener — no longer depends on Chrome re-applying suggested_key after updates

NewDocumentation page at web2md.org/docs — covers quick start, Obsidian export (with prerequisites), Send to AI, Batch Convert, Prompt Templates, and keyboard shortcut

Improved"View Docs" link in extension settings now points to web2md.org/docs

v0.4.0February 21, 2026

NewPrompt Templates — 4 built-in templates (Summarize, Translate, Key Points, ELI5) plus custom template CRUD

NewBatch Convert — paste URLs or select open tabs for bulk markdown conversion with progress tracking

NewSite Extractors — dedicated extraction for Reddit, Stack Overflow, GitHub Issues/PRs, YouTube (with captions), and Twitter/X

NewObsidian Integration — one-click 'Save to Obsidian' via obsidian:// URI protocol

NewNotion Integration — OAuth connect and export markdown as Notion pages

New3 new AI platforms — Perplexity, Grok, and DeepSeek added (6 total: Claude, ChatGPT, Gemini, Perplexity, Grok, DeepSeek)

NewKeyboard shortcut — Control+M (Mac) / Alt+M (Windows) for instant page-to-markdown conversion

ImprovedSettings redesigned with 3-column grid for AI platform selector

ImprovedContext menu shows keyboard shortcut hint

v0.2.5February 20, 2026

NewGA4 analytics monitoring system with user properties and conversion funnel events

NewEvent tracking for all user actions: convert, copy, download, send to AI, settings changes

ImprovedUser properties synced to GA4: plan type, login status, preferred AI platform

v0.2.4February 19, 2026

NewUnified preview card container — all cards wrapped in a single panel with shared action bar

NewCheckbox selection — select/deselect individual cards, action buttons show selected count

NewCard display duration setting — choose 5s, 15s, 30s, or Always in Settings

New"Always" mode persists cards to chrome.storage.local — cards survive page navigation

ImprovedBottom action bar replaces per-card Copy/Send buttons — cleaner UI for multi-card workflows

FixedPrompt prefix now reliably applied — fixed Zustand localStorage vs chrome.storage.local sync issue

v0.2.3February 18, 2026

NewPreview card — persistent card on page after conversion with Copy and Send to AI buttons

NewMultiple preview cards stack with batch Copy All / Send All actions

NewSmart context menu — "Copy selection as Markdown" when text is selected, "Copy page as Markdown" otherwise

NewRight-click conversion also shows preview card on the page

ImprovedPreview card auto-dismisses loading toast for cleaner UX

FixedPreview card injection uses chrome.scripting API — works reliably without page reload

v0.2.1February 18, 2026

FixedReddit conversion now uses JSON API — bypasses Shadow DOM for reliable post and comment extraction

ImprovedRight-click context menu shows specific error when free limit is reached instead of generic failure message

v0.2.0February 18, 2026

NewSend to AI — one-click paste converted Markdown into Claude, ChatGPT, or Gemini

NewCustom prompt prefix — prepend instructions before sending content to AI

NewAuto-send to AI option — convert and send in a single click

NewRight-click "Send to AI" context menu for selected text

v0.1.9February 18, 2026

NewReddit support — extracts post title, body, images, and up to 30 comments

NewDedicated Reddit extractors for both new Reddit (shreddit-*) and old Reddit (.expando/.md)

v0.1.8February 17, 2026

ImprovedChanged free tier from 20 conversions/day to 20 total conversions — upgrade to Pro for unlimited

FixedFixed review prompt not appearing after conversion — now re-checks after each successful conversion

NewAdded Blog link to navigation bar

v0.1.7February 16, 2026

NewAdded 12 popular platforms to Supported Sites: Reddit, X/Twitter, LinkedIn, Quora, MDN Web Docs, CSDN, Sspai, 36Kr, TechCrunch, YouTube, Notion

ImprovedEnriched use case descriptions with more specific site references across all 6 categories

ImprovedUpdated all 10 language translations (EN, ZH, JA, KO, ES, FR, DE, PT, RU, AR)

v0.1.6February 15, 2026

FixedFixed text truncation on JS-hydrated sites by extracting complete content from CMS embedded data

FixedFixed nested JSON.parse("...") parsing for CMS preloaded data (Substack, Ghost, WordPress SSR)

ImprovedAlways-compare content strategy: DOM vs JSON-LD vs embedded body_html — picks the most complete result

v0.1.5February 15, 2026

FixedGeneralized CDN proxy URL handling — restores original image URLs from any CDN proxy

FixedFixed URL-encoded image URLs not being decoded correctly in CDN proxy patterns

NewSmart image extraction from data-attrs JSON, data-orig-file, and data-lazy-src attributes

NewCDN proxy link cleanup — <a> href URLs with /image/fetch/ are now restored to original URLs

NewJSON-LD articleBody and embedded body_html fallback extraction

v0.1.4February 14, 2026

New11-language extension support (EN, ZH-CN, ZH-TW, JA, KO, ES, FR, DE, PT-BR, RU, AR)

FixedFixed text truncation caused by overly broad CSS selectors removing inline elements

FixedFixed Substack CDN image URLs by extracting original URLs and removing signatures

ImprovedTwo-phase DOM cleanup to protect inline elements during content extraction

ImprovedAdded content selectors for Substack and Ghost platforms

FixedEliminated all hardcoded English strings — full i18n coverage

v0.1.2February 10, 2026

NewChrome Web Store listing with optimized description and screenshots

NewIn-app review prompt after 3 conversions

FixedUpdated Chrome Web Store links to actual extension URL

v0.1.1February 7, 2026

NewOG social preview image for link sharing

NewGA4 event tracking and MDX blog system

NewSEO infrastructure with sitemap and llm.txt for GEO

NewFavicon using .MD brand logo

ImprovedDomain migration to web2md.org

FixedFixed CORS issues with consistent domain usage

FixedRedesigned magic link email with dark theme

v0.1.0February 1, 2026Initial

NewInitial release — full-stack Web2MD application

NewOne-click web to Markdown conversion via Chrome extension

NewSmart content extraction powered by Mozilla Readability

NewUser account system with Google login and magic link authentication

NewPro subscription with unlimited conversions

NewRight-click context menu: copy page or selection as Markdown

NewToken counting, word count, and reading time estimation

NewAuto code language detection for code blocks

New10-language website (EN, ZH, ES, FR, DE, JA, KO, PT, RU, AR)

NewExtension i18n with English and Chinese