We2ima

This release rebuilds the packaging and update pipeline and improves the import experience: distribution moves from a zip archive to a proper installer, the app no longer extracts files to the system temp directory at every launch (ending antivirus false positives and file locks that broke startup), automatic updates now silently run a full installer, new adaptive open-mode handling guides you when WeChat's default-browser option mismatches (with a one-click mode switch), a Slow Import Mode notice was added, and Fast Import is now a persisted preference covering scheduled and Full Auto tasks — and is significantly faster (measured ~150 articles/hour daytime single-directory, ~95 articles/hour dual-directory); installing to a non-system drive keeps your data safe across OS reinstalls. It also includes everything from the never-released v1.0.4 cycle.

• ✨ Installer-based distribution. No more temp-folder extraction at runtime, eliminating antivirus false positives and locks; faster, more stable startup
• ✨ Automatic updates now download the full installer and upgrade silently, then relaunch the app
• ✨ New install-integrity self-check: after an OS reinstall, the app checks the desktop shortcut, uninstall entry, and startup item at launch and offers one-click repair
• ✨ The default install directory now prefers drive D: to protect payment credentials and the database from OS reinstalls
• ✨ New adaptive open-mode handling: when WeChat's "open web pages with the system default browser" option mismatches the current open mode, a dialog guides the fix — switch to "Open Articles via URL" with one click and keep importing (the current article is not wasted), or fix the option and let the article be re-opened and re-detected automatically; unattended tasks are handled safely without dialogs
• ✨ New Slow Import Mode notice: starting manually without "Fast Import" explains the current ~30 articles/hour slow mode and how to switch (Fast Import reaches ~95-150 articles/hour,different speeds during night/day), with the option to cancel and enable Fast Import first
• 💡 Fast Import is now a persisted global preference: your choice is saved permanently, survives restarts, and applies to manual, scheduled, and Full Auto tasks
• 💡 Fast Import is significantly faster: redundant fixed waits removed in favor of event-driven waiting, and the IMA extension version check now runs only once per task; measured ~150 articles/hour (single-directory) and ~95 articles/hour (dual-directory) in daytime (~125 / ~75 at night) — nearly double the previous speed; night-batch default is now 10 articles per batch with a 2-minute rest
• 💡 Chrome extensions are now bundled in the installer — ready on first launch, no download needed
• 💡 Installing to a non-system drive (e.g. D:) is recommended so payment credentials and database files survive OS reinstalls
• 🐛 Fixed duplicate We2ima desktop shortcuts after installing via the setup package: the install self-check now recognizes both the public and user desktops, repairs to the public desktop, and removes the leftover duplicate from the user desktop automatically on first launch after upgrading
• 📝 Upgrading from v1.0.3 or earlier (zip version)? Follow the migration guide to keep your scan data, import history, and license — no re-activation needed on the same PC
• 📝 Starting with v1.1.0, the in-app updater of older zip versions no longer receives new releases; please download the installer from this page to migrate
• 📝 If you have never set a database password, your old database cannot be opened after an OS reinstall (the Windows encryption key is rebuilt with the OS) — set one in Settings in advance
• 📝 Windows SmartScreen may show "Windows protected your PC" for the installer (no code-signing certificate yet) — click "More info → Run anyway"; a single UAC prompt during an automatic update is expected
• 📝 Since v1.1.0, only Google Chrome is supported (must be installed, no need to set as default); Edge/QQ/360 Speed/Quark are no longer adapted

This local-client feature and bug-fix release adds 30-day tag trends, a read-marker publisher ranking, automatic updates, a new Quarterly plan, plan discounts with pay-time price verification, and favorites tag-list export, unblocks the “Open Articles via URL” import path for non-Chrome default-browser setups (short links copied from the built-in WeChat viewer, eliminating duplicate ima imports), reorganizes the read-marker statistics page with sidebar navigation, improves read markers, first-run onboarding, scan/import workflows, and Settings stability, and strengthens import and scan pause handling, incremental-scan resume, interruption recovery, IMA scaling and browser-zoom compatibility, extension-version checks, and Chrome main-window detection, and fixes the packaged build's startup/exit temporary-directory errors, while adjusting paid-plan prices and quotas.

• ✨ Added an in-app automatic update entry. When a new version is available, users can view update status and start the automatic update flow from the client
• ✨ Added browser extension update support. The automatic update flow can handle both the IMA import extension and the read-marker extension
• ✨ Added first-run theme onboarding. New users can preview and choose a UI theme after accepting the agreement
• ✨ Added a review entry for the first-use readiness guide in the About page
• ✨ Added startup subscription reminders based on current quota, unused activation codes, and trial status
• ✨ Added a first-run “User Research” dialog that collects how users discovered the app (channel, platform, content form) with an optional email subscription for update notifications
• ✨ Added runtime compatibility checks for the IMA extension. Before each real IMA click cycle, We2ima checks the extension version currently running in Chrome; outdated, missing, malformed, or unresponsive extensions stop safely before any click and show an upgrade-guide entry
• ✨ Added an experimental IMA DOM-click mode that can be enabled manually in Settings. Real-mouse clicking remains the default
• ✨ Added “30-Day Tag Trends” from Tag Overview, with a Top-20 chart of articles added over the last 30 days and an animated monthly ranking across the last 12 months. The monthly ranking supports play/pause, month-by-month stepping, a month slider, and 0.5×/1×/1.5×/2× speeds. Charts use the current app theme and open offline in the system browser
• ✨ Added a “Publisher Rankings” view to the read-marker statistics page, showing the Top 20 publishers by newly read articles in the last 30 days and an animated monthly ranking across the last 12 months. Playback controls match the tag-trend page, manual month changes keep the paused state, and the system “reduce motion” preference is respected
• ✨ Added a proactive read-marker extension upgrade reminder. When an outdated read-marker extension is detected running, the app shows at most one reminder per launch, with direct access to the official upgrade guide covering backup, in-place replacement, and reload steps
• ✨ Added plan discount display and pay-time price verification. Subscription plan cards can show the strikethrough list price, the discount (e.g. 20% off), and the sale price; clicking “Pay” re-checks the latest server catalog first, and if prices or discounts changed, the app refreshes the plan cards and asks you to confirm before paying again; the charged amount is always calculated by the server from the discounted price
• ✨ Added a new Quarterly plan tier. The subscription page now shows 6 plan cards instead of 5, including the new Quarterly plan (¥29.9 / 5,000 imports / 90 days), with full support for Quarterly activation codes, subscription reminders, and pay-time price verification
• ✨ Added an “Export WeChat Favorites Tag List” button on the Home page. Scanned user-defined favorites tags can be exported to a txt file (one tag per line) with an automatic date-time stamp in the file name, saved to a location of your choice
• ✨ Added automatic default-browser detection and adaptation. On every launch, We2ima checks the system default browser: if it is not Chrome, the “Open Articles via URL” mode is enabled automatically (no need to change the system default browser — you only need Chrome installed); if it is Chrome, the mode stays off
• 💡 Adjusted paid-plan prices and quotas (effective 2026-08-21): Monthly ¥9.9 / 1,000 imports / 30 days; Half-Yearly ¥68 / 15,000 imports / 180 days; Yearly ¥128 / 35,000 imports / 365 days; Permanent ¥188; Pay-Per-Use ¥0.012 per article. The 50-article free trial, the 5-articles-per-day free tier, and the ¥5 Full Auto unlock remain unchanged
• 💡 Improved read-marker matching to reduce missing read overlays on some newer WeChat Official Account articles
• 💡 Improved read-state persistence so progress is recorded more promptly while reading, reducing state loss when the browser is closed directly or exits unexpectedly
• 💡 Improved the first-use readiness guide with clearer information about setup effort, task duration, import scope, collection organization, and paid plans
• 💡 Improved scan and import date-range input. If dates are entered in reverse order, the client now corrects them automatically and continues
• 💡 Improved Home page date-range refresh. When scan data changes, default date ranges refresh after returning to Home without overwriting user-edited dates
• 💡 Improved scan recovery by removing the old “continue from last stop” radio option and consolidating recovery into the unfinished-scan task entry
• 💡 Improved Settings language and theme synchronization so selection fields stay aligned with the actual configuration
• 💡 Improved extension directory selection with a new in-app directory picker that replaces the system folder dialog
• 💡 Improved the first-run release-notice dialog by removing duplicated version details and linking to the official download page for the full changelog
• 💡 Improved IMA coordinate mapping for Chrome page zoom levels of 50%, 80%, 100%, 125%, and 150%, plus common Windows display-scaling configurations on a single monitor. If the coordinate space cannot be verified, the task now stops safely instead of guessing a click position
• 💡 Improved experimental DOM-click timing by waiting about two seconds after the target directory is scrolled into view and centered before clicking
• 💡 Improved the IMA extension upgrade dialog so it is centered relative to the main application window
• 💡 Improved import-task pause responsiveness. Fast and slow imports now honor Alt+P during article loading, IMA status polling, extension-popup triggering, and pre-click waits; remaining waits resume afterward without counting paused time against timeouts
• 💡 Improved first-open rendering and ordering for tag-trend charts so the 30-day chart is visible immediately and sorted by descending new-article count
• 💡 Improved the read-marker statistics page layout. The top tabs are replaced by a dark-blue sidebar with four sections — Overview, Read Articles, Publisher Rankings, and About; the sidebar keeps the “Local data” marker to stress that statistics stay on the device; the new About section explains the feature scope, counting rules, local-data privacy, and the minimum extension version; on narrow screens the navigation opens via a menu button
• 💡 Improved visual consistency between the read-marker statistics page and tag trends. Both trend pages now share locally bundled brand fonts, the real app icon, capsule tabs, unified chart cards, and the brand-blue palette, with a current-month watermark on monthly trends; fonts and icons are embedded locally, work offline, and load no external resources
• 💡 Improved transparency of the 30-day tag trend page. The subtitle now states the tag scope (top tags by new articles in the last 30 days), the exact statistics window dates, and that deleted articles are excluded, making it easier to reconcile with manual counts in the Statistics list
• 💡 Improved resuming of interrupted incremental scans. When continuing an unfinished incremental scan, We2ima reuses the per-tag baseline dates saved when the task started, scanning incrementally past the interruption point to fill the gap instead of re-scanning the whole category from the top; chained resumes after repeated interruptions are supported
• 💡 Improved the automatic-update download experience and status feedback. When a new version is detected, the installer is pre-downloaded in the background while the status bar shows staged messages (“downloading / ready / download failed”); the update dialog shows live download progress and speed; failed downloads are retried automatically, and if all attempts fail the app directs you to download manually from the official website, with guidance to try alternative download sources when the download is slow
• 💡 Improved how article URLs are obtained in “Open Articles via URL” mode. URLs are now copied as official WeChat short links from the built-in WeChat viewer (“More → Copy Link”), located entirely by UI automation with no position calibration required; the short links match the URL form used by mobile and browser imports, so ima deduplicates correctly by URL. If the precondition is not met (the WeChat option “Open web pages with the system default browser” is still enabled), the task stops with step-by-step guidance and can be resumed from the breakpoint afterward
• 🐛 Fixed read overlays not appearing correctly on some WeChat Official Account articles because publisher information was matched too narrowly
• 🐛 Fixed read-marker status loss when users closed the entire browser directly
• 🐛 Fixed possible duplicate counts or inconsistent states when read-marker status could be written from multiple paths
• 🐛 Fixed stale scan/import date-range display that could become inconsistent with the current database state
• 🐛 Fixed Settings language and theme selectors showing stale values or mismatched display names in some cases
• 🐛 Fixed startup exceptions that could occur when language callbacks ran before the status bar was fully initialized
• 🐛 Fixed Tk scheduling exceptions that could occur when background threads requested UI updates during startup
• 🐛 Fixed the extension-folder Browse button leaving the app permanently unresponsive
• 🐛 Fixed breakpoint loss and inconsistent fast/slow-mode window recovery after abnormal import interruption (e.g. WeChat crash); unified to three-attempt window recovery with a structured task outcome and per-article safe checkpoints
• 🐛 Fixed the community QR code never appearing in the first-use readiness dialog when the payment server was unreachable, and fixed long QR-section titles overflowing and being clipped
• 🐛 Fixed scaled or unmoved real-mouse clicks on IMA directory items when Chrome page zoom or Windows DPI changed the coordinate space
• 🐛 Fixed valid viewport geometry at 50% Chrome zoom being rejected, which could prevent directory targeting and incorrectly show an IMA extension upgrade warning
• 🐛 Fixed normal article titles containing “Chrome DevTools Protocol” or “ima.copilot” being mistaken for auxiliary windows and failing to import
• 🐛 Fixed Chrome auxiliary windows such as “Translate this page?” being selected as the article window, which could cause coordinate failures or a false extension-upgrade warning
• 🐛 Fixed fast or slow imports continuing to open and click the IMA extension after Alt+P; Alt+Q still stops immediately while paused
• 🐛 Fixed automatic imports, fast imports, resumed tasks, and Full Auto tasks potentially creating read sessions and polluting read history or statistics. Automated opens are suppressed by default, while explicit scrolling, clicking, or key activity after pause or cleanup can still start a session
• 🐛 Fixed 30-Day Tag Trends selecting the Top 20 tags by all-time article count, which pushed tags with recent additions out of the list and made the 30-day chart incomplete. Tags are now selected by new-article count within the last 30 days, tags without recent additions no longer occupy slots, and both tabs share the same tag set
• 🐛 Fixed scan pausing often appearing to fail because holding Alt+P slightly longer let keyboard auto-repeat toggle pause back and forth; the pause hotkey now fires on key release, so each physical press toggles exactly once
• 🐛 Fixed Alt+P/Alt+Q being completely unregistered when a scan was restarted from the same dialog after the previous scan finished or was stopped
• 🐛 Fixed delayed pause response during scanning (pausing only took effect after the current batch of articles finished), Alt+Q not stopping during fast date jumps, and a resume race that could skip window-state recovery
• 🐛 Fixed interrupted incremental scans re-scanning the entire interrupted category (such as “Links”) from the top when resumed, and fixed interruptions caused by abnormal WeChat exits being misrecorded as completed
• 🐛 Fixed the IMA extension being falsely reported as “unresponsive/outdated” in some scenarios, which aborted import tasks (a typical case: the article was already in the default directory and only the target tag directory still needed importing)
• 🐛 Fixed auto-update signature verification failures leaving the “Upgrade automatically” button permanently disabled with no error shown in the UI or logs; the update signing tool now builds the manifest exactly as the server does, and verification failures are shown explicitly in the update dialog
• 🐛 Fixed “Open Articles via URL” mode failing to find Chrome when the default browser was not Chrome (especially per-user Chrome installs), which prevented articles from opening and caused them to be wrongly marked as skipped after retries
• 🐛 Fixed the right-click menu popping upward when the target article was in the bottom rows of the favorites viewport, causing the “Copy Link” click to miss and the article to be wrongly marked as skipped; such articles are now scrolled to the upper-middle of the viewport first, while articles at the very end of the list are handled after scrolling to the bottom
• 🐛 Fixed duplicate entries in the ima knowledge base when the same article was imported through different channels: URL-open mode previously captured long links whose form differed from the short links saved by mobile imports, defeating ima’s URL-based deduplication; official WeChat short links are now used consistently, eliminating duplicate imports
• 🐛 Fixed articles already present in the ima knowledge base not being recognized in URL-open mode, which caused the ima menu to be clicked twice (removed, then re-collected); URL-open mode and click-open mode now share the same already-imported detection pipeline, and already-imported articles are skipped with zero menu clicks
• 🐛 Fixed the guidance dialog for unmet URL-open preconditions potentially popping up twice
• 🐛 Fixed articles whose publisher name contains spaces failing to locate in URL-open mode, and silently disappearing with no retry and no record; locating failures are now retried within the retry budget, and final failures appear in the task report and are marked as skipped
• 🐛 Fixed the viewer menu missing the “Copy Link” item when opened before the article finished loading, which caused the viewer to be opened twice per article; the menu now opens only after the page has loaded, and a menu missing the item is closed and retried within the same viewer
• 🐛 Fixed “Scan ima Import Directory List” failing when the default browser was not Chrome — the article link is now obtained via the built-in WeChat viewer and forcibly opened in Chrome; also fixed the feature erroring out when the top favorites article was not an Official Account article — a prompt that auto-dismisses after 5 seconds is now shown and the next articles are tried automatically until one usable for retrieving the directory list is found
• 🐛 Fixed the packaged build occasionally failing to start with a “Failed to load Python DLL” error, and occasionally showing a “Failed to remove temporary directory” error on exit; the app now automatically purges stale extraction directories older than 24 hours at startup (directories still in use are skipped safely), waits for background components to fully release file handles on exit, and prevents new leftovers through a single-instance constraint on the main app and disabled UPX compression for the launcher — the failure rate now converges with use instead of growing
• 📝 Note: After browser extensions are updated automatically, restart the browser or open chrome://extensions/ and refresh the updated extensions to enable the new extension features
• 📝 Note: If automatic update is unavailable or the standalone updater is missing, users can still download the full release package from the official website and update manually
• 📝 Note: Date-range order is corrected automatically for scan and import flows, but empty or invalid dates still need to be completed or corrected manually
• 📝 Note: Read-marker recovery for abnormal browser exits is improved, but the browser extension still needs to be enabled and allowed to run on article pages
• 📝 Note: Upgrade and reload the read-marker extension to 0.1.1. Older versions show an upgrade notice and do not follow the new import-source isolation rules; see the read-marker extension upgrade guide for step-by-step instructions (Chinese)
• 📝 Note: The extension-directory picker only accepts existing local directories; network paths, mapped network drives, and symbolic links are not supported. Copy the extension to a local fixed drive first if needed
• 📝 Note: Existing users do not need manual migration; resume checkpoints are auto-compatible with the new recovery logic
• 📝 Note: v1.0.4 requires the IMA optimized extension currently running in Chrome to be version 1.1.3.1 or later. If you replace only the main executable, follow the IMA extension upgrade guide to reload or install the new extension
• 📝 Note: IMA coordinate adaptation currently supports single-monitor configurations only. Multi-monitor setups, cross-screen movement, mixed per-monitor DPI, and changing the window/display environment during a task are not supported
• 📝 Note: Experimental DOM clicking is disabled by default. If the DOM path fails, We2ima does not fall back to a real-mouse click, preventing duplicate clicks on the same directory
• 📝 Note: The Alt+P pause hotkey for scans and imports now takes effect when the key is released; normal quick presses feel the same as before, and holding the key no longer toggles pause repeatedly
• 📝 Note: After upgrading, the database gains the fields required for scan baselines automatically — no manual migration needed. Unfinished scans left over from older versions still resume with the previous full re-scan behavior; incremental scans interrupted after the upgrade resume with the new baseline logic
• 📝 Note: Plan discounts are configured on the server; the client shows the strikethrough list price and sale price, the charged amount is the discounted price calculated by the server, and the payment dialog shows only the final amount due
• 📝 Note: The Quarterly plan requires We2ima v1.0.4 or later; older clients cannot recognize or activate Quarterly activation codes — please upgrade first. Adjusted plan prices always follow the payment server's live catalog
• 📝 Note: Do not run a scan while an import task is running — both tasks would register the same pause/stop hotkeys and their states can conflict; stop the import task before scanning
• 📝 Note: If your default browser is not Chrome, you do not need to change the system default browser — just install Chrome and, in WeChat “Settings → General Settings”, uncheck “Open web pages with the system default browser”. We2ima enables the “Open Articles via URL” mode automatically at launch; if the precondition is not met, the task stops with guidance
• 📝 Note: “Open Articles via URL” mode no longer requires “Copy Link” menu-position calibration — viewer menus are located directly by UI automation using element names; any previously completed calibration data is kept as a fallback and does not affect usage
• 📝 Note: If a v1.0.4 installer downloaded before 2026-08-31 shows an unavailable “Upgrade automatically” button or fails to download the update package, please download the latest installer again from the official download page and update manually
• 📝 Note: The packaged main app no longer allows multiple instances — a second launch shows “We2ima is already running. Please check the system tray.” and exits automatically; closing the main window minimizes the app to the system tray instead of quitting, so double-clicking the desktop icon again shows this message, which is expected behavior
• 📝 Note: If antivirus real-time scanning (e.g. 360, Windows Defender) blocks the app from starting, add the main executable to the trust list — for Defender, add We2ima.exe under “Virus & threat protection → Exclusions → Processes”; for 360, add the installation directory and the main executable to the trusted zone

Adds WeChat Official Account article downloading and read-status marking, while improving upgrade compatibility, read-marker stability, release notices, and packaging completeness.

• ✨ Added WeChat Official Account Article Download with Markdown and MHTML output for local archiving
• ✨ Added WeChat Official Account Read Marker with read indicators in WeChat Favorites article lists
• ✨ Added read-marker statistics and management pages for reading time, publisher stats, and overview data
• ✨ Added a v1.0.3 release notice dialog for new and upgraded users
• 💡 Improved upgrade compatibility by backfilling new settings and creating backups before required database migrations
• 💡 Improved build and packaging configuration so new modules, extension assets, dependencies, and version metadata are included
• 🐛 Fixed read-marker records being incorrectly created when import tasks automatically open WeChat articles
• 🐛 Fixed legacy main-database upgrades that could fail when indexes were created before missing fields were added
• 🐛 Fixed repeated creation of database migration backup files
• 📝 Markdown and MHTML output may vary depending on article structure, image loading, and network conditions
• 📝 WeChat Read Marker works with a local service and Chrome extension. Reading status is recorded and displayed locally on the user's machine.And only applicable to the old version of WeChat (version v3.9.x).
• 📝 During automated import tasks, We2ima avoids marking auto-opened articles as read. If the task is paused and the user continues reading the current article, it can be treated as manual reading.
• 📝 When upgrading to v1.0.3 for the first time, the app may create settings or database migration backups to protect older user data.

Focuses on local-client import efficiency, import start mode flexibility, debugging workflows, WeChat article import stability, favorites scanning reliability, full reprocessing reliability, and first-run onboarding.

• ✨ Experimental Fast Import reduces per-article waiting time while keeping necessary waits and retries
• ✨ Browser Open Mode lets you test article locating and browser opening when IMA import is disabled
• ✨ Browser Open Mode configuration separate history, statistics, and debug quota support
• ✨ First-run onboarding readiness dialog with setup reminders and a Feishu user community entry
• 💡 Improved import start settings by separating import range from processing behavior, including From Position or Date Range with Process All
• 💡 Improved Only New and Resume From Last Stop interactions by moving resume into a separate recovery button and dialog
• 💡 Improved scheduled-task and full-auto-task start mode display so new and legacy settings reflect the actual execution strategy
• 💡 Optimize the instability of the structure import of some articles on WeChat official account, and reduce the probability of import failure
• 💡 Optimize Chrome article window recognition to reduce automatic window title localization failures
• 💡 Improved favorites scanning so video/image-only tags finish normally and continue to later tags
• 🐛 Fixed transient import failures being incorrectly saved as permanent Skip states, allowing later retries
• 🐛 Fixed “From Top (Process All)” mode so all historical article states can be retried, and maintain the normalization of the database status
• 🐛 Fixed From Position + Process All and Date Range + Process All so these combinations can be represented and executed reliably
• 🐛 Fixed legacy scheduled tasks, full-auto tasks, and resume records that could otherwise display inconsistently or run with the wrong import range
• 🐛 Fixed Only New edge cases with reverse direction, date-range cursors, and Browser Open Mode that could stop too early or use the wrong direction
• 🐛 Fixed full-auto subtasks saving empty or non-positive limit values that could otherwise fail later during execution
• 🐛 Fixed the import task report dialog possibly being hidden behind browser windows; it now opens topmost by default
• 🐛 Fixed the issue of calculating the hierarchy of "batch quantity" in multi Tag import mode
• 🐛 Fixed the task interruption caused by mouse drift or abnormal movement during ima automatic positioning and clicking
• 📝 Browser Open Mode does not mean the article has been imported into the IMA knowledge base
• 📝 Resume From Last Stop is now a recovery action, not a normal saved start mode for scheduled or full-auto tasks

First post-release iteration focused on scan stability, date-range scanning, user experience and bug fixes.

• ✨ Date-range scanning
• ✨ Resume incomplete scans (one-click continue on remaining tags)
• ✨ Detailed import-task report on completion
• ✨ Lark (Feishu) community QR code on payment success page and seed-user activity
• ✨ Software version upgrade notification feature
• 💡 Scan delays configurable (day/night dual-tier)
• 💡 Optimization of shortcut key responses (Alt+P to Pause / Alt+Q to Stop) in scanning tasks
• 💡 Set page layout refactoring
• 💡 Slow-scan delays randomized
• 🐛 The issue of "custom shortcut keys" taking immediate effect
• 🐛 The issue of 'restoring default settings'
• 🐛 Fix 3 boundary issues in date-range scanning mode
• 🐛 Fix early-stop regression in incremental multi-tag scanning
• 🐛 Fix "Link" tag ima-folder badge display anomaly
• 🐛 Fix tasks incorrectly marked "scan complete" after abnormal WeChat exit

• Initial public release
• 6 import modes (skip processed, full, resume, position, date range, new only)
• Fast favorites scanning with deduplication
• Smart resume from breakpoint with pause/resume support
• Date range filtering
• Statistics dashboard and import history
• Bilingual interface (Chinese / English)
• Pure local operation, zero data collection