Adds clearly marked Fast, Professional and Ultimate local Image OCR entries for Chinese, Japanese and Latin scripts, reconstructs source rows and tab-separated table columns, and automatically falls back between verified model sources.
ImgIng / Changelog and roadmap
CHANGELOG · PUBLIC ROADMAPWhat shipped, what comes next, and why
A user-facing product record containing only changes people can use or perceive. Shipped capabilities are dated, future work is labelled separately, and deployment, operations and internal validation details stay out.
Multi-layer layout, editable multilingual text, professional filters and skin-aware beauty have shipped; next comes crop, graphics, selections and local edits.
RELEASE NOTES
Shipped
Every item below is a feature, fix or experience change users can already perceive; concepts and internal engineering details are excluded.
CURRENT
New local multilingual image OCR workspace
Image OCR is now available with a New badge from both the main task area and the first position in the side shortcut menu. Fast supports Chinese, English, German, Spanish, Portuguese and French; Professional and Ultimate add Japanese. Models use a public download source first, switch automatically to the matching site-hosted file when needed, and reuse verified browser cache entries. Results rebuild visual rows, paragraphs and table columns from text-box geometry. Copy and TXT preserve source rows and separate table columns with tabs, and copying shows an immediate success confirmation. Regular column runs become Markdown tables, while JSON retains raw and corrected text, coordinates and confidence. The right panel uses a compact layout preview with collapsible line proofreading. The current unified model excludes Korean and does not claim arbitrary handwriting or formula semantics as reliable capabilities.
SHIPPED
Animation Workshop reduces GIF and APNG delta size and improves multi-format defaults
GIF safely stores local changes and automatically clears full frames when a transparent outline changes, preventing ghosting. Animations that use exactly 256 colours also keep every real colour instead of sacrificing one for delta transparency. On the current 685×150, 134-frame sample, default GIF and APNG outputs are both about 1.17 MB and all 134 decoded frames remain pixel-identical to the source; the same APNG is about 6.55 MB with full frames. Colour-rich, high-frame-count material defaults to animated WebP, and multi-format ZIP export follows the selected delta- or full-frame mode.
SHIPPED
Faster video matting without changing the current quality path
After previewing and confirming the first video frame, full export now reuses the established high-quality first-frame matte and temporal state, continuing from the next frame instead of repeating the most expensive calibration. Frame processing avoids unnecessary image copies and repeated memory preparation, while progress and output previews update at a steadier pace for a more responsive long-video workflow. The Experimental Ultimate model, RVM ResNet50 hair calibration, ten first-frame refinements, five-frame temporal-memory cadence, scene-cut rebuilding, aspect-matched analysis size and final output pixels remain unchanged.
SHIPPED
AI cutout now starts local inference correctly on mobile WeChat and desktop
On first use, mobile WeChat and desktop browsers can now load and start local inference without stopping because a runtime address cannot be resolved or the inference backend fails after the model download completes. Mobile keeps the memory-safe Fast AI choice, while desktop retains Fast, Pro and Ultimate models. Models still use the public download source first, switch automatically to the site fallback when needed, and reuse completed browser downloads.
SHIPPED
PDF compression resumes exactly where you left it
PDF compression now keeps the current file, preset, preview mode and completed output when its window is closed. Reopening restores a valid preview instead of showing a broken file. The per-image decision list also uses all remaining space on taller displays and stays scrollable on shorter ones, so completed compression details are no longer cut off above an empty area.
SHIPPED
More scalable charts, tables and text in HTML-to-PDF output
HTML to PDF now prevents common shadows from turning whole charts, tables or text regions into images whenever the appearance can be preserved; complex effects without a proven equivalent remain unchanged. Lossless minimization prepares in parallel with page conversion and recursively merges identical images and transparency masks in the same write, so re-importing the result into PDF compression is no longer needed to reveal another structural reduction. On the current 14-page chart-heavy report, page count, geometry, backgrounds, watermarks and shadows stay unchanged; extractable text rises from about 1,297 to 2,716 characters, one automatic pass reduces the file from about 4.82 MB to 1.56 MB, and page generation falls from about 5.3 to 2.6 seconds. Savings vary by page content; Canvas, real images and effects that require raster output remain pixel content.
SHIPPED
Substantially smaller PDF-to-HTML files without reducing page fidelity
The local fixed-layout engine advances to 1.3. Repeated vector components, paths and presentation styles that can be shared safely are stored once across the document, while permitted embedded TrueType and OpenType fonts use a lossless WOFF container. The process does not lower image quality, rasterise pages, or remove or substitute permitted fonts; complex components without a proven equivalent remain fully expanded. On the current 15-page chart-heavy report, output falls from 12,719,382 bytes to about 4.75 MB while retaining all 15 pages, 29 embedded fonts, selectable text and every chart. Savings vary with the amount of reusable content in each PDF. The result remains one offline-ready HTML file with no viewer controls.
SHIPPED
Complete Spanish and Portuguese experiences
The language order is now Chinese / EN / Japanese / Korean / DE / ES / PT / FR. Spanish and Portuguese extend beyond Home to the image and video workspaces, PDF compression, PDF to HTML, HTML to PDF, task guides and release history. First visits recognize common es and pt regional variants from the browser’s complete preference list, while explicit locale URLs and manual choices always win. Wide layouts show full names and constrained layouts use short labels, keeping all eight choices on one line without covering the brand, theme control or tools.
SHIPPED
Online Design becomes a multi-page workspace
An Online Design project can now contain up to 24 independent pages, each with its own canvas size, background, images, editable multilingual text and filters. Canvas presets follow the actual dimensions and identify non-standard sizes as Custom. On-demand fonts repaint immediately after loading, and a font change affects only the selected text layer instead of shifting other text. Existing single-page image export remains available, joined by page-by-page ZIP and standalone high-resolution multi-page HTML. The open .imging format advances to v3 with every page while remaining compatible with v1 and v2; existing v3 project content can still be opened and edited. Chinese, English, Japanese, Korean, German and French provide the complete workflow.
SHIPPED
WeChat-safe mode and one memory guard for mobile AI cutout
WeChat, iPhone and Android browsers now check model support before downloading: Fast AI remains available, while Pro AI and Ultimate AI are clearly labelled desktop-only and explain why when selected instead of downloading high-memory assets first. Very large photos preserve their aspect ratio and use the current mobile device’s safe pixel budget without cropping or stretching. Compact preview, Professional Refinement, background composition, fast colour removal and final export share the same safe dimensions so no later step recreates an oversized canvas. Model cards, method controls, progress, preview and mask tools gain larger mobile type and touch targets. Images remain on-device, while desktop keeps every model and full-resolution output.
SHIPPED
Precise six-locale entry from the complete ordered browser preference list
Chinese, English, Japanese, Korean, German and French recognize their common regional and script variants. When the first browser preference is unsupported, the entry logic continues through the ordered list and selects the first supported locale; English is used only when none match. A language chosen by the user and an explicitly opened locale URL always take precedence, avoiding incorrect changes based on country, proxy network or later detection. Home and all five image, video and PDF workspace shortcuts share the same rule.
SHIPPED
PDF to HTML and HTML to PDF now use official status consistently
Home task cards, workspace launchers, workspace titles, feature introductions and task guidance no longer mark PDF to HTML or HTML to PDF as Beta. Chinese, English, Japanese, Korean, German and French stay aligned. The change promotes the current feature status without hiding fixed-layout reconstruction, browser compatibility or server-side Chromium conversion boundaries; Video Matting remains Beta.
SHIPPED
Multilingual text layers across image export, Online Design and Professional Refinement
Image compression and format export keep their draggable text watermark and add fonts covering Chinese, English, Japanese, Korean, German and French. Online Design adds independent editable text layers with content, font, size, colour, alignment, line height, opacity, scale, rotation, layer order and 2.5D effects. Professional Refinement after AI cutout adds the same directly draggable text-watermark layer. Chinese uses Alimama ShuHeiTi; English, German, French and other Latin languages use Noto Sans; Japanese and Korean use their matching Noto CJK fonts. Font files load and cache only when first selected. Preview, full-size export and professional output wait for the same font, preventing preview/export substitutions. The open .imging project format advances to v2 with readable text parameters while retaining v1 image-project compatibility, so hand-off does not require rasterising text. After HTML to PDF completes, the exact preview frame used for output is locked so source-page hover scripts cannot keep changing it; hidden Home decoration also pauses while a full-screen workspace is open.
SHIPPED
Complete Korean and French experiences across one six-locale system
The language order is now Chinese / EN / Japanese / Korean / DE / FR, using full labels on wide screens and short one-line labels on narrow screens. Korean and French are complete experiences rather than Home-only buttons: both include Home, image and video workspaces, PDF compression, PDF to HTML, HTML to PDF, twelve task guides and release history. Explicit locale URLs always win; first visits to Home or the five workspace shortcuts now match Korean and French browser preferences, while unsupported languages continue to use English. Opening or closing a workspace, pressing Escape and navigating Back all preserve the active language.
SHIPPED
Separate Chinese, English, Japanese and German choices with complete localized routes
The top bar exposes Chinese, EN, Japanese and DE as four separate choices with a visible active state. Narrow screens use short one-line labels without crowding the brand or theme control. Manual selection is stored locally and wins over automatic detection; a one-use query choice keeps switching reliable when storage is blocked and is removed immediately after reading. First visits to Home and the five workspace shortcuts match Chinese, Japanese or German browser preferences and otherwise fall back to English; explicit locale URLs are never hijacked. All four locales include workspace routes, task guidance, release notes and counterpart language links.
SHIPPED
PDF compression with two-axis DPI, inline images and content-aware structural cleanup
PDF compression continues to optimize the original document structure without an HTML reconstruction round trip. Image analysis aggregates horizontal and vertical placement independently across every use, allowing non-uniform placement to choose each target dimension safely. Decode arrays are applied per component, while colour spaces that cannot be proven equivalent remain untouched. Safely parsed BI / ID / EI inline images are promoted to standard Image XObjects and enter the same classification, effective-DPI and encoding race as ordinary embedded images. Content operators drive conservative removal of unused fonts, images, graphics states, colour spaces, patterns, shadings and properties; unknown operators or special streams keep their complete resource set. Fragmented page content streams merge only at safe lexical boundaries and only when smaller. Existing Flate streams also race a fresh lossless encode and change only when it wins. Output is re-read to verify page count, page boxes, selectable text and link semantics, with automatic fallback whenever verification differs or the complete file is not smaller.
SHIPPED
Workspace exits now reliably restore the home URL and active language
Fixed Online Design, Video Matting, PDF compression, PDF to HTML and HTML to PDF closing visually while leaving the tool shortcut in the address bar after import, preview or processing. Whether launched from Home or opened directly, X, Escape and browser Back now close the workspace and restore / for Chinese or /en/ for English. Refreshing then stays on Home instead of reopening the closed tool.
SHIPPED
HTML to PDF with true pages and generic lossless vector/font-resource minimization
HTML to PDF Beta keeps Import → Preview → Explicit Convert → Automatic Minimization → Direct Save. A single HTML previews immediately, while a complete folder can authorize relative CSS, images, fonts, scripts and data. Conversion settles document.fonts, images, charts and Canvas only after the user clicks Convert, then derives true geometry from CSS @page, fixed-layout per-page metadata or the complete rendered size instead of applying hidden A4. The automatic preset is now generically lossless: it does not recognize one chart library or SVG template, but promotes mathematically identical filled paths in PDF content streams to shared Forms while preserving each call’s transform, colour, alpha and clipping. Visible Chinese and Latin fonts remain intact; Type 3 glyph drawing is compacted to its original metrics only when every use is proven invisible and never clips text. Image pixels are neither resampled nor re-encoded by default; print 300 dpi, e-book and screen presets require an explicit choice. The output is re-read to verify page count, every page box, selectable-text operators and link semantics, with automatic fallback when anything differs or no smaller result exists. The final size and suggested filename are shown before direct saving, with no printer or PDF re-selection. During conversion, the top action keeps a fixed width and status label while detailed phases update only in the progress panel, preventing the right-side controls from flashing or reflowing. Fonts, images, dynamic charts and final page measurement are now prepared in an invisible isolated view, so the central source HTML preview is never reloaded from click through save and the first conversion no longer produces a full-page white flash or refresh-like jump. Instant preview now supports sandboxed single-file pages that rely on eval, WASM or Workers, and keeps an explicit opening state until the iframe has actually loaded instead of reporting an unrendered white view as complete. Font/image settling now reports real completed counts, elapsed seconds and an active progress signal while distinguishing local decoding from network downloads. Dynamic pages reuse the true dimensions measured during their first isolated render instead of loading all fonts and images again merely for measurement. Fixed-layout pages now use each metadata page size directly instead of deriving print height from an aspect ratio; page count and every page box are verified before saving to prevent trailing blank pages or geometry drift. Layout capacity is prepared ahead of use and safely reused between jobs, so small pages no longer repeat a fixed startup wait. When no web font or image needs decoding, conversion continues immediately and the progress text no longer suggests a font download.
SHIPPED
PDF to HTML with cross-version masks, colour and page-specific quality reporting
The local fixed-layout engine is now version 1.2. It fixes reversed gradients and missing JPEG images with SMask in real presentation PDFs, and adds one-bit ImageMask stencils, inline images, JPEG soft masks, common graphics-state Alpha and Luminosity masks, ICCBased fallbacks, indexed palettes, radial gradients, and standard static appearance streams for annotations and forms. FontDescriptor weight and style are retained, while soft-mask content and Type 3 glyphs are shared across pages. The bilingual quality report maps each issue code to exact page ranges; CMYK JPEG, custom Decode arrays, colour-key masks, JPEG 2000, special soft-mask transfer functions and matte premultiplication are no longer silently treated as equivalent.
SHIPPED
Video Matting with unified naming and a scoped lab-grade quality statement
Home launchers, the video workspace, bilingual task pages and footers now consistently use Video Matting. The feature is positioned as a lab-grade Beta: the MatAnyone2 ultimate tier delivered the best overall matting result on current representative internal samples among the comparable browser-side solutions evaluated. The wording also preserves the essential boundary that this is a scoped, provisional internal result rather than an absolute ranking of every clip, device or industry product. Existing /video and /video-editor/ URLs remain available.
SHIPPED
PDF to HTML with corrected gradient coordinates, smaller vectors and a seamless full preview
The local clean-page converter reads Type 3 CharProcs, shares identical SVG outlines across the complete document and keeps a transparent selectable-text layer. PatternType 1 tiling patterns and PatternType 2 axial shadings now use the absolute pattern matrix already supplied by the PDF instead of multiplying the active graphics matrix again. This restores chart gradients in Chrome / Skia presentation PDFs that previously collapsed to a dark solid. In a real 15-page test file, repeated Type 3 definitions fell from 3,625 to 1,012 and output size dropped from about 6.3 MB to about 4.38 MB. The report now explains that compressed PDF streams can still expand when represented as script-free, directly rendered SVG with selectable text. A Web Worker delivers page one first and continues the same parse for all remaining pages; the preview becomes the complete scrollable document automatically, and Generate reuses that result. Offscreen pages render on demand, closing unloads the document, and the original PDF remains an optional comparison link.
SHIPPED
PDF to HTML with fixed-layout pages, licensed fonts and no viewer chrome
The home page now includes a PDF to HTML Beta launcher plus /pdf-html shortcuts in both languages. The self-built local engine parses page boxes, rotation, content streams, graphics state, text matrices, ToUnicode, vectors, images, forms and links; ordinary pages become fixed-layout SVG with selectable text. TrueType and OpenType fonts are checked through OS/2.fsType and only permitted fonts are inlined; Restricted License and bitmap-only fonts are never copied. The single HTML contains only the CSS, SVG, images and fonts needed by the pages, with no external loads, PDF JavaScript, zoom, report, print or original-PDF viewer controls. Non-equivalent features such as JPEG 2000, mesh shading, text clipping and complex forms remain in the workspace per-page report for source comparison before saving. The result opens offline and uses a suggested filename when saved.
SHIPPED
One system save flow for exports, downloads and editable projects
Image conversion and compression, AI cutout, single and three-format animation output, Online Design images, open .imging projects, video background replacement and PDF compression now use one save flow. Desktop Chrome and Edge with the File System Access API show a system dialog with a safe non-overwriting suggested name and let the user confirm the destination; long design, project, animation bundle and video jobs choose that target before encoding. WeChat, iPhone, Safari and Firefox keep a compatible download fallback and clearly point users to browser downloads or the Files app.
SHIPPED
Reliable direct routes and memory-safe PDF queues
Opening /img, /pdf or their English equivalents now reliably launches the intended workspace. Batch PDF import no longer reads every file eagerly: each source is loaded only when processed, source bytes are released afterwards, original File objects are reused when compression cannot shrink them, device-aware file/count/total limits are enforced, and a Clear queue action explicitly releases retained results.
SHIPPED
PDF compression that re-encodes each embedded image at its measured DPI
A PDF compression workspace now sits below the video editor entry. The self-built engine parses xref tables, xref streams and object streams locally, walks each page’s content streams while tracking the graphics state matrix, and derives every embedded image’s real effective DPI; only images above the chosen preset are resampled, with Lanczos-3. Each image is classified by ImgIng’s already-calibrated content classifier, then JPEG, indexed-palette Flate, grayscale Flate and keeping the original compete and the smallest wins — with every decision listed. Writing out performs reachability collection, re-deflates uncompressed streams, strips private data and thumbnails, and repacks with object streams; the result is then re-read and reverted if the page count or any page box differs. Screen 72dpi, e-book 150dpi, print 300dpi and lossless-structure presets are available along with a target file size. PDFs and their images never leave the device and no extra runtime is downloaded. Fonts are not subset, so text-only PDFs gain little, and encrypted PDFs, JPEG 2000, JBIG2 and CMYK print JPEGs are left untouched.
SHIPPED
Professional watermark layers across conversion and compression
Enter Chinese, English or brand text, or add a transparent PNG / WebP logo. Position it directly on a compact canvas, resize and rotate with handles, then tune colour, opacity, placement and either a flat or soft 2.5D treatment. The layer is redrawn locally at final resolution, while disabled watermarks leave encoding, caching and source-preservation behaviour unchanged.
SHIPPED
Clearer launcher typography and mobile readability
Increased task titles, format descriptions, NEW / BETA badges, trust promises, expandable guidance and brand contact text, together with card height and spacing. The two-column mobile layout now allows long task titles to wrap to two lines without clipping, crowding or misalignment.
SHIPPED
Smaller format-aware defaults by content type
Retuned recommendations for regular and smooth photos, texture, graphics and soft alpha around the official Squoosh and codec baselines. Native browser JPEG retains extra headroom, WebP alpha stays lossless at 100, common phone-size AVIF images use a more efficient effort tier, and same-format compression still keeps the source whenever re-encoding is not smaller.
SHIPPED
Clear home hero and direct task launchers
Kept the existing two-column structure and full converter while changing the information order to task first, explanation second. Six launchers open conversion, compression, AI cutout, Animation Studio, Online Design and Video Background directly. Local processing, no-signup and no-watermark promises remain visible, with technical detail available on demand across English, Chinese and mobile layouts.
SHIPPED
Local WebP processing and mobile layout fixes for iPhone WeChat
WebP capability checks now validate the actual MIME type so WebKit cannot silently return PNG. When Canvas WebP encoding is unavailable, a self-hosted libwebp WASM encoder loads on demand across conversion, same-format compression, background removal, Online Design and animation export without uploading pixels. Mobile text scaling, header controls, mode icons and floating launchers now have stable sizing, baselines and safe-area spacing.
SHIPPED
Video background colour selection and instant preview fix
Fixed background changes hiding an already-matted frame and revealing the white source video. Colour, image, blur and background-mode changes now reuse the cached matte for instant recomposition without rerunning AI, while the control shows both a real swatch and its hex value.
SHIPPED
Professional filters, skin-aware beauty and open projects
Added per-image-layer and full-canvas scopes with 12 presets and 18 light, colour, skin-aware beauty and detail controls. Edge-preserving smoothing never expands transparent contours, alpha is retained pixel for pixel, large images use overlapping memory-safe bands, and .imging stores all parameters, original assets and a filtered preview.
SHIPPED
Video first-frame hair, aspect and blank-intro fixes
Fixed a fallback-mask payload bug and low-contrast hair loss in the experimental tier. RVM ResNet50 now calibrates the first frame before MatAnyone2 temporal memory; narrow gaps are closed in linear time without replacing the original soft alpha. Landscape, square, portrait and ultrawide video keep their aspect ratio, while blank intros output the chosen background and continue looking for a later subject.
SHIPPED
Unified light/dark editor themes and scrollbars
Online Design, Video Editor and Professional Refinement now share the site design tokens across toolbars, panels, controls, canvases, cards and exit confirmation. Each workspace also gets an in-editor theme control and matching scrollbars.
SHIPPED
Ultimate video model status overlap fix
Video model cards now use an adaptive two-column title and status layout. The long MatAnyone2 title and Cached for ratio state no longer overlap, while the description retains the full row width.
SHIPPED
English home launcher overlap fix
Online Design and Video Editor launchers now share an auto-flow container instead of separate fixed vertical coordinates, preserving stable spacing for wrapped English copy and responsive layouts.
SHIPPED
Bilingual image and video shortcut routes
Added /img, /video, /en/img and /en/video direct launchers. Chinese and English users can open the correct workspace directly from home or a task page, then keep the active language when closing it or navigating back.
SHIPPED
Video export size and bitrate fix
Fixed oversized MP4 and WebM output after background replacement. Video bitrate now adapts to the source, resolution, frame rate and alpha channel; compatible audio is copied directly and only necessary transcoding uses an adaptive audio bitrate.
SHIPPED
Online Video Editor Beta
Opened local video background replacement with two RVM models and aspect-adaptive MatAnyone2 landscape, square and portrait variants. Models use ModelScope CDN first with site fallback; original aspect ratio and audio are retained, with MP4 or transparent WebM export.
SHIPPED
Animated WebP ghosting fix and frame-optimization control
Fixed price and text ghosting caused by lossy animated WebP delta-frame blending. Restored the Delta frames / Full frames control and routed WebP, APNG, GIF and AVIF through their matching encoding paths.
SHIPPED
Frame scaling and arbitrary-angle image rotation
Shipped Keynote-style eight-way scaling and rotation handles, precise inspector angles and ±90° shortcuts. Preview, undo/redo, .imging project round-tripping and final image export all preserve rotation.
SHIPPED
Online Design and an independent multi-layer workspace
Shipped a local design workspace with up to 24 layers: multi-image import, move, scale and order, transparent or solid backgrounds, canvas presets, undo/redo and full-size local PNG, WebP or JPG export.
SHIPPED
Local multi-background compositing after AI cutout
Professional refinement supports transparency, a solid colour or up to 12 local background layers. The subject and every background layer can move, scale, position and reorder independently, with full-resolution local compositing on export.
SHIPPED
Complete English interface and localized model states
Shipped the complete English tool and eight English task guides, with consistent English buttons, operation prompts, help copy, AI model names, download progress and runtime states.
SHIPPED
Professional AI cutout, real export and large-image performance
Unified full-resolution alpha across preview, refinement and download. Added a true 1 px brush, multi-step undo/redo, Space-to-pan, black/white edge inspection and a fixed-divider Squoosh-style comparison.
SHIPPED
Three local AI tiers and fast colour removal
Added on-demand ISNet INT8, BEN2 FP16 and BiRefNet HR-Matting FP16 models. The fast path removes only similar colour connected to the image edge, preserving matching colour inside the subject.
SHIPPED
End-to-end animation and isolated workspaces
Completed compression, creation and frame editing for GIF, APNG, animated WebP and animated AVIF, including 100 ms–3 s timing, loops, playback direction, drag ordering and isolated workspace state.
SHIPPED
Capability boundaries and tool comparison
Documented live browser capability detection, device/server boundaries and a transparent comparison methodology for ImgIng, Squoosh, TinyPNG and cloud converters.
PLANNED · NOT SHIPPED
Next release: deeper local image editing
Build frequent refinement tools on the shipped multi-layer workspace, with preview/export parity as a release requirement.
Canvas and geometry
Crop, flip, free transform and additional composition guides.
Local editing
Brush, eraser, local adjustments and selections connected to AI masks.
Alignment and more layer styles
Guides, alignment, distribution and additional layer styles come next.
Professional history
Non-destructive operation history, multi-step undo/redo and original/result comparison.
Device-side export
Keep real encoded-Blob verification and export common formats locally.
Planning note: this scope is a public target, not a currently available feature. Every item must pass image-quality, memory, browser-compatibility and large-image performance checks; incomplete work will be delayed instead of shipped prematurely.
PRODUCT STANDARD
Roadmap principles
Every new capability must pass four product rules.
Local first
If a browser can complete the work reliably, image pixels stay on-device.
Results you can trust
Preview, comparison and download share the same real-pixel pipeline.
Large images remain usable
Zooming, panning and editing must stay responsive at high resolution.
Boundaries stay explicit
Planned, device-side, on-demand WASM and server features are labelled separately.
RESEARCH · NO COMMITTED DATE
Research pool
Directions without a committed date stay in research, separate from the next release.
Batch workflows and reusable presets
Apply one set of dimensions, quality, naming and export rules to many files.
Masks and finer selections
Extend the shipped multi-layer composition with selections, masks and more precise local control.
Colour management and professional metadata
Evaluate ICC, EXIF, HDR and cross-browser colour consistency.
More local models and plugin-style capabilities
Add only when download size, licence, speed and quality are acceptable.
Release and roadmap questions
The same clear answers for people and search engines.
Can I use the future features shown here now?
No. Multi-page design, editable multilingual text, professional filters and skin-aware beauty are shipped, while crop, guides, selections and local editing remain planned. Only the Shipped section describes current capabilities.
Will the local image editor upload images?
The shipped Online Design text and filter tools keep assets local. The goal is to keep future crop, guides, brush, selections and common-format export in the browser too. Any server-only operation will be labelled before use.
Why publish target dates instead of invented version numbers?
ImgIng currently follows continuous delivery, so dates are easier to verify. Semantic versions can be introduced when the formal release process needs them.
Can the roadmap change?
Yes. Browser support, memory, large-image performance and export parity can change scope; updates will remain visible on this page.
Reviewed by the DataDance product and engineering team ·