Nouveautés

Added

• Public pages prerendered in every language, with one URL per language (/fr, /es, /it, /de)
• Page titles, descriptions, canonical URLs, language alternates, schema.org data, sitemap.xml, robots.txt, llms.txt
• "Try with a demo video" on the landing page
• Transitions at cuts: quick fade or fade to black
• Texts over the video (titles, notes) with the thumbnail title styles
• Background blur behind the person (MediaPipe selfie segmentation)
• Voice noise reduction (RNNoise) at export, with "listen to 8 s of the result"
• Save a project as a .lcv.json file and open it on another computer
• Unit tests (Vitest) of the engine and end-to-end tests (Playwright) of the key flows

Changed

• A reconnected video with the same name and size is accepted without confirmation
• Visitors are sent to the pages in their language

Added

• Speed change for the words selected in the transcript

Changed

• Dragging on the waveform selects a part (then Cut or Speed) instead of cutting it at once
• Cut and Speed are never greyed out: without a selection they explain how to make one

Added

• Speed changes: speed up or slow down a part (0.25× to 4×) from the Start / End marks, pitch kept in preview and export
• Thumbnail maker: hold ‹ › to scroll through the frames, ← → keys (Shift: 1 s)
• Auto edit button in the landing page mockup

Fixed

• Site header overflowing (flag-only language picker, icon install button, menu below 1200 px)

Added

• Thumbnail maker: best moments suggested, big title with highlighted words, 4 styles, JPG / PNG for YouTube, Shorts and square
• Video volume up to 300 % with a limiter, mute, clip leveling
• Project rotation in 90° steps (preview, export, face tracking)
• Auto edit asks whether someone speaks in each clip, with automatic voice detection
• Clips without voice: no transcription or subtitles, only real dead air is cut

Changed

• "Music" tab renamed "Sound"
• Silence cuts computed clip by clip
• Music ducking ignores clips without voice
• Export uses a peak limiter instead of hard clipping

Fixed

• Auto cut chopped videos without speech (street, nature) into pieces

Added

• Selected cut editor on the timeline: start / end nudges (±0.1 s), listen, remove
• Delete removes the selected cut, Escape deselects it

Changed

• Wider, visible cut edge handles
• An adjusted silence cut becomes manual and is kept by live re-detection

Fixed

• Drawing a cut on the waveform did nothing (drag-to-seek swallowed the drag)
• Inspector content clipped on the right (face tracking row)
• Transcript selection buttons cut off in a narrow panel
• Timeline crushed in short windows and the two-column layout
• Colored edges on the auto edit button

Added

• Auto edit: one click for transcription, fillers, silences, framing, zoom and subtitles; undo in one step
• Manual cut with Start / End / Cut buttons (keys I, O, X)
• Transcript correction: retype a word or a whole sentence (Enter)

Changed

• More compact preview on phones
• Audio uses the first decodable track (iPhone spatial audio ignored)

Fixed

• Reopened project stuck on the waveform loader after reconnecting a clip
• Audio analysis errors now show a retry instead of an endless loader
• Editor wider than the screen on phones

Added

• Automatic spoken-language detection (Whisper language ID, majority vote)
• "Transcribe everything again" and "Transcribe the new clips" buttons
• Auto-cut from the transcript, rhythm presets, cuts updated live
• Preview quality (360p / 540p / 720p) with a "light preview" badge
• Timeline zoom buttons, "Fit" and Ctrl + wheel
• Language picker in the editor
• Mobile editor layout with a bottom tab bar
• Styled dropdowns, sliders and scrollbars

Changed

• Resolution and frame rate moved to the export dialog
• Smart zoom also works without cuts (on sentences, else every 4 s)
• "Follow the face" starts face tracking by itself
• Silence detection: smoothing, hysteresis, short noises ignored
• Waveform scaled on the voice, not on the loudest sound

Fixed

• Timeline zoom stretched the whole editor
• Automatic language mode transcribed other languages into English
• French elisions split by Whisper ("j 'ai")
• Inspector content clipped on the left

Added

• Multi-clip projects: import several videos at once, reorder, add or remove clips
• Unreadable files are skipped and reported instead of blocking the import
• Legal notice, privacy policy and terms of use pages (5 languages)
• Language picker with flags and three-way theme switch
• Mobile navigation menu
• Page transitions and scroll animations

Changed

• Reconnect screen lists every clip with its own status
• Server access logs disabled

Added

• PWA: offline mode, install button, "open with" for video files, update toast
• App version in the footer and in-app changelog page
• Docker image (nginx, COOP/COEP, strict CSP) and Docker Compose for Coolify
• README with screenshots
• MIT license

Added

• Waveform timeline (Wavesurfer.js) with draggable, resizable cuts
• WYSIWYG canvas preview: skips cuts, plays ducked music
• MP4 export with WebCodecs + Mediabunny, streamed to disk when supported
• Undo / redo and keyboard shortcuts

Added

• Output formats: 16:9, 9:16, 1:1
• Silence auto-cut with adaptive threshold
• Whisper word-level subtitles: karaoke, clean, boxed
• Text-based editing and filler-word removal
• Face auto-framing (MediaPipe) and punch-in smart zoom
• Background music with auto-ducking

Added

• Media worker: probe, 16 kHz mono audio extraction in RAM, waveform peaks
• AI worker: Whisper via Transformers.js (WebGPU, WASM fallback)
• Self-hosted ONNX Runtime and MediaPipe binaries

Added

• Zero-upload import with the File System Access API, file input fallback
• Local database with Dexie (IndexedDB)
• Auto-save with RxJS debounceTime
• Crash recovery: reconnects the video file on reopen

Added

• Angular 22 app: standalone components, zoneless, signals
• Landing page
• Native light/dark theme ( prefers-color-scheme ) with manual toggle
• i18n: EN, FR, ES, IT, DE