202654 releases
November8 releases
Improved- Fixed the available speech speeds for ElevenLabs Multilingual v2 in text voiceover.
- Added a link to the marketplace product video translation guide on the English video translation page.
Improved- Improved timeout handling for transcript and translation correction. When a brief network issue interrupts an audio download, VoxSail retries the completed result without repeating synthesis.
Improved- Reduced the wait before dubbing starts, avoided repeated checks of voice-cloning reference audio, and sped up audio post-processing.
Improved- Improved resource allocation across tasks and fixed voiceover draft saving, restoring dubbing for individual segments, and language estimates with dubbing disabled. Refined upload and editor messages.
Improved- Fixed an issue that could stall progress when many tasks were queued. Completed results now update promptly, helping voiceover tasks and queued work continue.
Improved- Improved reference audio handling and error feedback for very short speech segments, making voiceover preparation more reliable. Added complete English and Chinese status messages for vocal separation tasks.
Improved- Improved resource allocation and queuing for voiceover tasks, reducing cases where an invalid task blocks later tasks. Clearer feedback is provided when processing capacity is insufficient.
Improved- Uploads now flag videos without audio tracks and files that exceed your plan limits earlier. In the editor, you can save speaker changes and move existing voiceovers.
- Updated the default translation model and the English and Chinese help guides, including upload limits and project workflows.
October11 releases
Improved- Improved number and model-name pronunciation in voiceovers, reference audio handling, and task recovery. Added clearer notices and purchase restrictions for frozen credits, plus a guide to testing translations.
Improved- Improved Chinese subtitle segmentation, task recovery, and video playback compatibility on Windows. Refined project names, failure messages, and admin list refresh, and added a supported languages guide.
Improved- Added a website introduction in Chinese and English, refreshed blog covers, and improved the layout of article headings, lists, and tables.
Improved- Improved portrait video editing with expanded previews and resizable panels. Subtitle edits are saved before translation and voiceover actions, and nicknames now support a single character with clearer length guidance.
Improved- Improved large-file upload efficiency and manual retry when source files expire. Storage usage is now available in Usage Analytics, with better preparation and recovery for lip sync on high-resolution videos.
Improved- Improved fallback and recovery when some voiceover requests are rejected, with better queueing and shared voiceover capacity to reduce repeated generation and keep billing consistent.
Improved- Repair individual sentences after content screening interrupts part of a voiceover, then export again. Improved screening retries, failure messages and credit protection when a repair fails.
Improved- Choose a voice before entering text and adjust supported native speaking rates. Improved long-form narration, upload recovery and model availability, with clearer no-speech and expired-cover states.
Improved- Improved upload queues, file limit and URL import feedback, made editor retranslation and export checks more reliable, and added workspace announcements and product introduction videos in Chinese and English.
Improved- Improved long-audio editor waveforms and resumable mobile uploads, enabled cleanup of all terminal projects, and added a bilingual subtitle format guide.
Improved- Improved task queue concurrency controls, upload entry behavior, and media processing reliability.
September28 releases
Improved- Fixed conflicts when linking an existing phone number after WeChat sign-in, and improved recovery from connection failures in speech recognition and video dubbing.
Improved- Improved language switching, footer navigation, and blog typography; fixed sharing images and added homepage links to relevant guides.
Improved- Updated the Chinese brand name, improved referral reliability, and added a guide to terminology consistency in course translation.
Improved- Improved subtitle sentence boundaries and timing, with cleaner endings for dubbed audio.
Improved- Improved subtitle boundaries, source waveforms, and dubbing loudness matching to reduce clipped speech, timing drift, and volume jumps.
- Improved original-voice cloning, short-line expression, speaker notes, and recovery from failed tasks.
- Unified media uploads and improved editor preview generation, sign-in continuity, and retry reliability.
Improved- Improved demo video loading, playback, and retry controls on mobile devices and embedded browsers.
Improved- Improved upload continuity after sign-in, WeChat phone binding, upload compatibility, duplicate import prevention, and feedback for large downloads.
Improved- Improved homepage loading speed with better-sized images and faster above-the-fold content.
Improved- Improved subtitle timing submissions when audio downloads are slow.
Improved- Improved search titles and descriptions, and redirected the retired audio translation entry to the combined video translation page.
Improved- Improved WeChat sign-in and signing back in, Chinese homepage switching, and upload guidance.
Improved- Improved speech timing and speaker segmentation accuracy, with easier speaker management and customer support access.
Improved- Fixed exports failing when retained original sounds, such as shouts, intentionally have no synthesized voice track.
Improved- Failed tasks allow up to 3 manual retries. The retry icon remains available afterward and shows a reminder to contact support.
- Delete subtitles by speaker while preserving source-audio mute ranges, with improved saving and undo.
- Fixed repeated synthesis after delivery and editor saving issues after background tasks finish.
Improved- Improved speaker separation and subtitle timing for multi-speaker videos, with more reliable recognition and segmentation.
- Improved dubbing transitions, retained source-audio details, and editor preview synchronization to reduce skipped or overlapping speech.
- Improved upload compatibility and error messages so submission issues are easier to understand.
Improved- Faster initial editor loading, with voices loaded on demand and fewer duplicate profile requests.
Improved- Improved editor loading speed and fixed repeated loading and flickering in some older projects.
Improved- Improved editor buffering: playback pauses together while audio loads and resumes when ready, reducing skipped speech.
Improved- Smoother endings for dubbed audio segments and improved reference handling for original voice cloning.
Improved- Fixed original voice cloning failures in single-speaker mode while preserving detected speaker identities.
Improved- Faster credit estimates and task submission, with less waiting and reconfirmation when prices change.
- Clearer upload states and a fresh workspace after reloading, without recovery prompts for unsubmitted files.
- Change the speaker for subtitle segments, with improved previews and source voice reference handling.
Improved- Improved search indexing by removing redirect-only documentation URLs from the sitemap and linking guides directly to their canonical pages.
Improved- Added a distinct cover for each blog post and refined the visual presentation of article lists and detail pages.
Improved- Added independent subtitle and dubbing track movement and per-segment volume controls, with improved source and dubbing audio previews.
- Improved mobile upload, project results, and editor layouts, with easier access to retry failed tasks.
- Fixed persistent loading in some video previews and improved translation retry and audio segmentation reliability.
Improved- Improved text correction speed after transcription to reduce processing wait times for long audio and video.
Improved- Fixed concurrency capacity remaining occupied after cancellation and prevented cancelled projects from re-entering the processing queue.
- Improved dubbing timeline boundaries and automatic speed adjustments to reduce segment overlap and improve transitions.
- Improved sign-in guidance after guest uploads while preserving uploaded content and the current workflow.
Improved- Improved original-voice reference reuse and historical-data compatibility for more reliable segment regeneration in the editor.
- Improved target-language menu expansion and collapse behavior for smoother multilingual selection.
- Refined editor risk messages and duplicate prompts so important action feedback is clearer.
New- Expanded original-voice cloning and multilingual voice selection for more source scenarios.
Improved- Improved task queuing, resource scheduling, cancellation, and retry reliability for long-running jobs.
- Improved subtitle editing, partial retranslation, segment regeneration, and re-export feedback.
- Improved media preview autoplay and lightweight feedback interactions with fewer layout shifts and duplicate prompts.
- Fixed duplicate verification-code submissions and back-navigation issues for smoother workspace access.
August7 releases
Improved- Improved context handling for continuous dialogue and editor retranslations, producing more natural and accurate transitions.
- Improved reference selection for original-voice cloning across speakers and short-source scenarios.
- Unified workspace layouts, icons, and loading transitions for smoother navigation and first-load experience.
New- Added more natural dual-reference voice cloning to better preserve a speaker's voice and expression.
Improved- Improved segmented voice preview reliability in the editor to reduce occasional silence, delays, and overlapping playback.
- Improved speech recognition, subtitle translation, and voice generation reliability with more accurate handling of anomalous content.
- Unified queue and processing states for final exports, with clearer project progress and failure details.
New- Added a dedicated video dubbing workspace with standard and professional multilingual workflows.
Improved- Improved language selection, original-voice guidance, and upload previews for video dubbing.
- Subtitle style and position now carry over from the last setup, with presets matching the final output more closely.
Improved- Improved project status accuracy during final video export with clearer progress and processing stages.
Improved- Improved segment mapping reliability for subtitle and dubbing translation to reduce failed tasks.
New- Added a lip-sync workspace with uploads, URL imports, and reuse of completed videos.
- Improved account redirects and result feedback after Stripe payments.
Improved- Improved source and target language selection, with Mandarin and Cantonese clearly separated.
- Projects now prefer original filenames for clearer browsing and management.
- Improved subtitle size and style consistency across languages.
New- Added multilingual voice-cloning showcases across several content types.
- Added more voices and multilingual previews for text-to-speech.
Improved- Improved language switching and comparison across video showcases.
- Refined the text-to-speech creation and delivery flow.
- Improved Explore voice availability during brief network interruptions.
This page highlights meaningful user-facing changes. Minor adjustments and internal maintenance are not listed individually.