Changelog

Free trial for everyone — every cloud feature opens without a redeem code. 50 credits per device, same point system as codes, with a one-time human check on first use.

• Added Free trial: AI image, audio transcription, TTS, and voice cloning are all open by default — 50 credits per device
• Added One-time quick human verification on first use, then quiet for 24 hours by default
• Added Top bar shows "Free credits: N" before a code is entered, with a buy-code prompt once you run out
• Added Trial toggle, per-device quota, per-IP device cap, and verification TTL are admin-tunable on the fly
• Improved Trial and redeem code share the same units and behavior; failed runs automatically refund credits

Full PDF toolchain on the canvas (Merge / Split / PDF→Images / Images→PDF); upload nodes get drag-to-reorder, PDF first-page thumbnails, and click-to-preview.

• Added Four new canvas nodes — Merge PDF, Split PDF, PDF to Images, Images to PDF. Runs locally in your browser, no credits used.
• Added Upload node accepts PDFs and auto-renders the first page as a thumbnail
• Added Drag-to-reorder for multi-file uploads — images, video, audio, and PDFs
• Added Click any thumbnail to preview — images zoom up, audio/video play inline, PDFs open in an embedded reader
• Improved Upload node size adapts to the media — wider for video, tight for audio, A4 ratio for PDF
• Improved TTS voice previews toggle play/pause directly — instant feedback, no spinner
• Fixed Preview animation occasionally flickered for large images or SVGs
• Fixed Files can now be removed or reordered on an upload node after it has already run
• Fixed Single-node runs now also block when credits are missing, instead of sending empty requests

Four AI image nodes land on the canvas (Upscale / Remove Background / Erase / Hero Poster), with a unified visual pass on parameter panels.

• New Four AI image nodes on canvas — Upscale, Remove Background / White Background, Erase (paint mask inside the node), and Hero Poster (multiple outputs per run)
• New Refreshed segmented switch across the app — three sizes, rounded / pill shapes, gentle bounce animation
• New Unified section headers across node parameter panels for cleaner scanning
• Improved AI nodes show an "AI" badge next to their name; the vectorize engine switch also marks the AI option
• Improved ASR model, TTS emotion / scene / speed / voice category — all switched to button groups with live descriptions
• Improved Watermark node panel re-laid out; single/batch switches on the four tool pages (image / audio / video compress + vectorize) are unified
• Fixed Cloud image results occasionally failed to display due to cross-origin issues
• Fixed Upload node staying busy after being emptied; small / portrait images being stretched

Pro vectorize engine on the canvas, nodes that auto-fit to media, and vectorize engine / cross-origin fixes.

• New Image vectorize gains a Pro engine on the canvas — switch local / Pro in a single node, with Pro billed per redeem code, sharing the same engine as the tool page
• New Canvas nodes auto-fit their height to the uploaded / produced media aspect ratio (animated); clearing the resource smoothly restores the default height, and single-file previews get a close button
• Fixed Fixed Pro vectorize failing to fetch the result SVG cross-origin — the server now returns the SVG text alongside the result, so the client no longer needs a cross-origin fetch
• Improved Vectorize Worker client extracted into a shared module; the canvas node (single instance) and tool page (Worker pool) reuse one protocol, removing the duplicate implementation

Full responsive support for the canvas workflow, a unified custom scrollbar, and two new canvas nodes: Text to Speech and Voice Clone.

• New New canvas nodes — Text to Speech and Voice Clone — sharing the synthesis core and voice library with the tool pages; text / reference audio come in via connections, audio comes out
• New Responsive canvas workflow: side panels become drawers on narrow screens, single-finger pan with a box-select mode toggle on touch, and an adaptive collapsing toolbar
• New Unified custom scrollbar component (AppScrollArea) — overlay style, consistent across Mac/Windows — adopted across all scroll areas
• Improved AppInput supports a leading icon and the node palette search gets a magnifier; the minimap renders thumbnail-less nodes as solid color; the crop node drops its dark background when empty
• Fixed Fixed overlapping left input ports on multi-input nodes (Voice Clone) that sent links to the wrong port; task frames now clip to the canvas area

Unified UI component library, in-browser local speech-to-text, server-side security hardening, and major upgrades to batch queues and the Pro vectorizer compare view.

• New Speech-to-Text adds a Local Privacy Mode: in-browser Whisper (WebGPU-accelerated), zero upload, no redeem code — with separate download/transcribe progress and a stop button
• New Site-wide UI component library — AppButton / AppInput / AppSlider / AppAlert / IconButton — plus dedicated 404 / 500 error pages
• New Pro vectorize now supports wheel zoom and drag-to-compare original vs SVG, matching the local engine
• Improved Image / audio / video batch queues: newest on top, common settings pinned above the queue, an Add Files button once files exist, and the size delta merged into the done status line
• Fixed Server hardening: per-IP daily rate limit on paid proxy endpoints, fail-closed auth, and timing-safe key comparison
• Fixed Fixed the invisible Back button on the 404 page; video queue ETA now shows an mm:ss countdown
• i18n Filled in tool tags, ASR model names and more in both languages; rewrote the Privacy Policy and About pages to match the current product

Launched the AI Image Toolkit and PDF Toolkit, added cloud speech-to-text / text-to-speech and voice cloning, and redesigned the home page.

• Launch AI Image Toolkit — upscale / cutout / erase / poster (Recraft engine), with results mirrored to our own R2 CDN
• New Voice cloning — Qwen-TTS voice replication and synthesis; text-to-speech with 49 voices plus static auditions
• New PDF Toolkit — merge / split / PDF-to-image / image-to-PDF with drag-to-reorder thumbnails, fully client-side
• New Pay-per-use cloud speech-to-text / text-to-speech, and a unified MaxAudioPlayer component
• Improved Home page redesign — variable-based card colors, AI-toolkit positioning copy, and purchase guidance via WeChat
• Chore Completed bilingual copy across the site and shifted SEO toward the AI-toolkit positioning

Full-site SEO upgrade, new zoom/confirm/batch workflow in the SVG vectorizer, and easier navigation between tool pages.

• New SVG vectorizer: fixed-size compare viewport with cursor-anchored wheel zoom, range 0.5x ~ 8x
• New SVG vectorizer: confirm button in both single and batch modes — uploads no longer auto-process
• New New RelatedTools section: the four main tool pages now cross-link at the bottom (Image / Audio / Video / SVG)
• Improved Added complete SEO fields to all six media tool pages: keywords / ogDescription / twitterTitle / twitterDescription
• Improved SVG vectorizer's scene-preset area now uses the same frosted card style as the image compressor
• Removed Removed the top-bar Docs link and the footer's Download Client block

Image compression now supports both single and batch workflows, with a video-aligned batch workspace and real stage-based progress feedback.

• New Image compression now switches between single-file and batch modes within the same page
• New Batch compression adds queue management, retry, per-item download and ZIP export
• Improved Batch workspace now matches the video compressor layout, with shared settings expanded by default and progressive queue/actions visibility
• Improved compress-sdk-plus now reports real image compression stages so the UI can show decoding, rendering, extracting and encoding progress

Full i18n coverage across tool pages. Fixed locale-aware navigation and hero theme colors.

• i18n Speech-to-text page fully localized: template strings, model labels and Groq error messages
• i18n Video compressor fully localized: single-file mode, batch mode, settings panel, summary report and action buttons
• i18n Home page tool card format tags (TTS / Voice Clone) now use i18n keys
• Fixed Home page tool card links changed from NuxtLink to LocaleLink to preserve locale prefix on navigation
• Improved CompressPageHero title and description colors now cascade through CSS variables to match each page's theme

Performance and stability improvements.

• Improved Page performance and build optimization
• Maintenance Dependency updates and bug fixes

Audio compression and speech-to-text are live. Batch download fixes.

• New Audio compressor with MP3, AAC, FLAC, OGG and WAV support
• New Speech to text with on-device Whisper and Groq / Together AI cloud engines
• New Together AI transcription via URL API
• Fixed Batch ZIP download now packages Blobs directly to avoid fetch-blob URL issues
• Removed Dropped unused together-ai and groq-sdk dependencies to reduce bundle size

Unified UI. Tool pages now share the home layout.

• New Image and video compressors now use the shared PageLayout
• New Added About, Privacy, Contact and Changelog pages
• Improved Video compressor blob animation now sweeps left-to-right
• Improved Darker, rounded backdrop for video thumbnail progress bar
• Improved Video size input now has clear focus styles and placeholder
• Fixed Aligned global CSS variables with the home page theme

Smoother video preview. Batch mode reliability fixes.

• Improved Handle-thumb preview now syncs to hover position precisely
• Improved extractSingleFrame for previews
• New Centered thumbnail progress bar
• Fixed Switching files now cancels the previous job and clears caches
• Fixed Better error handling when batch jobs are cancelled
• Fixed Use a cancelled flag instead of AbortController for reliable cancellation

Landing page launched. Video queue shows file size.

• New Landing page showcasing all available tools
• New Video compression queue now displays file size

MaxMedia launches with image and video compression.

• Launch Image compressor with JPEG, WebP, PNG and AVIF support
• Launch Video compressor with MP4, MOV, AVI and WebM support
• Launch Single and batch processing modes
• Launch Deployed on Cloudflare Pages