Speech Recognition & AI
Bilingual Mode (New)
VoicePing now supports real-time bilingual speech-to-text — speak in two languages during the same meeting, and both are transcribed and translated simultaneously without manual switching. Key highlights:- Sub-200ms latency for language detection and transcription
- Works with Japanese-English, Chinese (Simplified)-English, and Chinese (Traditional)-English pairs
- No manual language switching required — the system detects which language is being spoken automatically
Shorter Utterance Accuracy
Reduced hallucination issues with short utterances (e.g., brief responses like “yes” or “そうですね”) that could previously trigger incorrect transcriptions. Improved detection thresholds and filtering for more reliable short-phrase recognition.Speaker Diarization Improvements
- Improved speaker identification model with better accuracy for non-English languages
- New speaker merge feature — manually combine incorrectly split speakers in your transcript
- Better accuracy for non-English speaker separation
New Features
Transcript Keyword Search
Search across all your meeting transcripts by keyword. Results highlight matching text and scroll directly to the relevant section. Features include:- Full-text search across transcript titles, summaries, and spoken content
- Works across all languages
- Japanese IME-friendly input handling
- Paginated results with matched text snippets
Custom Phrase Auto-Extraction (Preview)
Early preview of an automated pipeline that identifies words your ASR model frequently misrecognizes, suggesting custom phrases to add. Full version with PDF/document upload launching in February.Listener Mode Transcription History
Listeners can now save transcription and translation logs locally in their mobile browser. Messages reset when the host session changes. Hosts can enable or disable this feature from the share options.Simplified & Traditional Chinese Support
VoicePing now fully supports both Simplified Chinese (zh-CN) and Traditional Chinese (zh-TW) across:- Web app and mobile app interfaces
- Notification emails
- Pricing and plan descriptions
- Documentation links (
manual.voiceping.net/zhandmanual.voiceping.net/zh-tw)
Data Intelligence Enhancements
- Sales Analysis: Talk ratio, filler word detection by language
- PM Dashboard: Word cloud, sentiment analysis, disagreement tracking, and speak ratio
- Supports Japanese, English, Simplified Chinese, Traditional Chinese, and Korean
Subtitle & Caption Improvements
- Shorter minimum duration: Dubbed files can now be as short as 30 seconds (previously 60 seconds)
- Multiple file upload: Select and upload multiple video files at once
- Direct text editing: Click directly on subtitle text to edit — no more pencil icon. Auto-saves on focus leave
- Real-time progress: Live status updates for upload, transcription, translation, embedding, and dubbing with progress percentages
- Processing notification: “We will email you” message now displayed for subtitle embedding and dubbing jobs
- Thumbnail fix: Audio-only MP4 files (no video stream) now correctly show a grey placeholder instead of failing
Meeting Log & Download Improvements
- Total participant count included in downloaded meeting logs
- PDF download: New lightweight PDF format — file size reduced from 446KB to 82KB for the same transcript with full CJK font support
- Improved file naming: Meeting start time added as prefix for easier sorting
- Translation dictionary files: Now use language-specific filenames instead of a generic name
Mobile App Updates
- v0.5.10 released for both iOS and Android
- Bilingual mode available on mobile
- Fixed message duplication during mobile transcription
- Fixed transcript language corruption in bilingual mode when AI summary was enabled
- Same language pair bug fixed: Correct languages now shown when identical pairs are selected
Bug Fixes
- Fixed CPU leak in desktop app caused by CSS animation running in an infinite loop after ending a session
- Fixed Italian language default: New users were incorrectly assigned Italian as their default language
- Fixed layout corruption when changing languages in the web app
- Fixed avatar/image loading issues in certain browsers (Brave, Sidekick)
UI/UX Improvements
- Bottom bar responsive layout: Action buttons no longer hidden on smaller laptop screens
- Subtitle download redesign: Direct buttons for Video, Subtitle, and Dubbing downloads with color-coded status
- Left-side scrolling fix: Data intelligence menu consolidated into a dropdown
- File upload size limit (25MB) now clearly displayed in the upload UI
- Meeting log selection: Clickable anywhere on the meeting title row
- Virtual office setting properly hidden in transcription mode