Changelog

Keep track of the latest features and improvements in Scribis.

v1.1.14

September 7, 2026
  • Qwen 3 TTS Memory Optimization: Resolved high memory usage during batch text-to-speech tasks by reusing tokenizers and speaker embeddings efficiently.
  • Upgraded Core AI Engine: Updated llama.cpp to version b10830 for improved performance and operational stability.
  • Backend Waveform Peak Extraction: Shifted Wavesurfer audio waveform peak calculation from frontend to backend for faster UI loading and reduced browser resource usage.
  • Automated Audio Noise Reduction: Added noise reduction support that automatically cleans up audio right after voice recording finishes.

v1.1.13

September 6, 2026
  • Smart Sensitive Data Redaction (Mosaic): Video editor now automatically detects and applies mosaic blurring over sensitive info like API keys, email addresses, and URLs.
  • Keyframe Support for Mosaic Blurring: Mosaics in the video editor now support keyframes, allowing you to easily track moving subjects or adjust blur areas over time.
  • Interactive Chapter Markers on Progress Bar: Add chapter markers directly to the video editor progress bar, with dynamic adjustment based on video length and playhead position.
  • Expanded AI & Speech Models: Added support for Breeze-ASR-26, MOSS-Transcribe-Diarize, Cohere, and Gemini 3.8 Flash models for expanded choice and enhanced transcription.
  • Upgraded YouTube AI Chat & Vision Mode: YouTube Chat now supports running local AI models alongside Gemini, introducing a Vision mode that optimizes token usage intelligently.
  • Right-to-Left (RTL) Layout & Subtitle Support: Full support for Right-to-Left languages (e.g., Arabic, Hebrew) across the user interface and subtitle rendering.
  • Automated Speech-to-Text Quality Re-Verification: Batch TTS workflows can now re-check audio quality via Speech-to-Text, allowing quick re-generation of lines with low accuracy.
  • EU AI Voice Watermarking Toggle: Added a compliance toggle for TTS audio watermarking in accordance with EU August 2026 AI guidelines.
  • Long Task & Resource Allocation Fixes: Resolved freeze issues caused by automatic memory release during long-running model tasks, and optimized dynamic library loading for smoother operations.
  • Enhanced Speech-to-Text Accuracy & Completeness: Fixed missing character issues in Qwen3-ASR and improved cross-platform accuracy across macOS architectures.
  • Video Subtitle OCR Re-Adjustment Fix: Resolved an issue where secondary adjustments to OCR detection boxes inside the video editor failed to register.
  • Fixed known issues and improved overall system stability.

v1.1.12

September 1, 2026
  • 16:9 to 9:16 Timeline Aspect Ratio Templates: Added timeline templates to easily adapt horizontal video layouts into vertical formats for short-form video platforms.
  • Preset Subtitle Templates: Introduced basic subtitle style presets to help you apply beautiful subtitle typography with a single click.
  • YouTube Download Quality & Format Options: Download YouTube videos with flexible format options and resolution choices (e.g., 1080p, 720p) via yt-dlp.
  • Quick Action Buttons in AI Chat: Added handy AI Chat shortcuts, including Bullet Points, Generate FAQ, and Create a Mindmap for instant content generation.
  • Auto-Matched AI Chat Language: AI Chat response defaults now automatically sync with your selected application interface language (i18n).
  • macOS x64 (Intel) Beta Build: Added testing build support for Intel-based macOS systems (x64).
  • Fixed known issues and improved overall system stability.

v1.1.11

August 30, 2026
  • Extended YouTube Subtitle Extraction: Support for auto-generating subtitles on longer YouTube videos without existing captions.
  • Resizable UI Panels: Added draggable splitters allowing you to freely adjust panel widths to customize your workspace.
  • Refined Scrollbar Styling: Visual enhancements to scrollbars across the application for a cleaner interface and smoother scrolling.
  • File Link Support for Batch Uploads: Batch importing folder files now supports referencing via file links without duplicate copies, conserving your disk space.
  • Real-Time Streaming Subtitles & Translations: Subtitle extraction and translation for YouTube videos now stream live, letting you view results instantly without waiting for processing to complete.
  • YouTube Mini Player Visual Layer Fix: Resolved an issue where the YouTube mini player in the video editor was covered or overlapped by other elements.
  • Comprehensive Tutorials with Auto-Translation: Added extensive guide tutorials that automatically translate according to your selected app language.
  • Block-Based Markdown Editor for Summaries & Notes: Introduced a block-based Markdown editor in AI summaries and notes for effortless note-taking and rich content formatting.

v1.1.10

August 28, 2026
  • Apple Intelligence & macOS Native Translation Support: Integrated Apple Intelligence and native macOS translation engines for faster, seamless multi-language translation.
  • Enhanced YouTube Player Controls: Added favorites, category management, and "Play Next" functionality for YouTube videos to better organize and stream content.
  • Speed & Audio Mixing Tracks: Introduced dedicated Speed and Audio Mixing tracks in the video editor, allowing flexible speed adjustments as well as volume control and muting for the original video audio.
  • AVFoundation Compatibility Fixes: Resolved crashes and errors caused by unsupported media formats in macOS AVFoundation.
  • Text & Subtitle Preset Previews: Save custom text styling configurations (such as font size and formatting) under custom names, and hover over saved presets to instantly preview text styles before loading.
  • Cloud ASR & Smart Voice Recording: AI Voice Assistant and voice-to-text features now support Groq, Gemini, and ElevenLabs Cloud ASR, with Voice Activity Detection (VAD) and automatic 10-minute timeout transcription.
  • Sample Projects & Interactive Testing: Added a wider variety of sample projects and test playgrounds to help users get started quickly.
  • Fixed known bugs and improved overall system stability.

v1.1.9

August 25, 2026
  • Regional Compliance Adjustments: Optimized OpenAI availability based on locale settings to ensure regional compliance.
  • Low Disk Space Warnings: Added intelligent low disk space alerts to prevent video rendering and export failures.
  • Streaming Media Architecture: Implemented a streaming-based media loading system to significantly reduce memory peak usage when processing high-resolution videos, delivering smoother performance.
  • MCP Endpoint Playground: Added an interactive API testing area within MCP endpoint settings to conveniently try out and test API endpoints.
  • Fixed known bugs and improved overall system stability.

v1.1.8

August 24, 2026
  • Mac App Store Stability Fixes: Resolved specific compatibility and runtime issues unique to the Mac App Store build.
  • Language Selection Improvements: Enhanced language handling for Speech-to-Text and Text-to-Speech; automatically defaults to English whenever an unknown or unrecognized language is encountered to ensure uninterrupted workflow.

v1.1.7

August 24, 2026
  • Custom Cloud API Support (BYOK): Added integration for Groq and ElevenLabs API keys, offering fast and flexible speech-to-text transcription.
  • Smart VAD Timing Correction: Automatically refines start and end timestamps via Voice Activity Detection post-transcription, eliminating overly extended ranges caused by background noise.
  • Waveform Splitting & Speaker Colors: Directly add cut points on the audio waveform, with automatic color-coding for distinct speakers.
  • Direct Inline Subtitle Editing: Edit subtitle text directly in place without needing to open separate pop-up windows.
  • Expanded Typography & Subtitle Styling: Added a broader selection of fonts and fine-tuned styling controls for cleaner visual presentations.
  • Transparent Video Export (ProRes 4444): Export subtitles with an alpha transparent background in Apple ProRes 4444 format for seamless video editor overlays.
  • YouTube Mini Player: Added a floating mini player with dedicated volume controls, allowing you to listen to video audio translated into other languages while editing subtitles for enhanced learning and productivity.
  • Expanded Keyboard Shortcuts: Introduced a comprehensive suite of hotkeys to accelerate your editing workflow.
  • Improved Mixed-Language Alignment: Enhanced accuracy for transcript matching and synchronization on mixed Chinese and English audio.
  • AI-Powered Smart Segmentation: Leverages LLMs to automatically re-segment sentences and optimize natural phrasing based on word timestamps.
  • Full Transcript Export: Easily export full verbatim transcripts for documentation and note-taking.
  • Magnetic Timeline Snapping: Subtitle start/end points within 200 ms now automatically snap together and drag seamlessly for effortless timing adjustments.
  • Fixed known issues and improved overall app stability.

v1.1.5

August 18, 2026
  • Upgraded llama.cpp to the latest version.

v1.1.4

August 18, 2026
  • macOS Media Processing Optimization: On macOS, MP4 and MOV formats now leverage native system processing for audio demuxing and subtitle embedding, while WebM formats continue using the browser engine for maximum performance and compatibility.
  • Flexible Frame OCR Selection: Hard subtitle OCR cropping is no longer restricted to frame 0. You can now navigate to any video frame where subtitles appear to set the bounding box area.
  • Hugging Face SHA-256 Model Verification: Added SHA-256 checksum verification for downloaded Hugging Face models to guarantee file integrity and security.
  • YouTube Assistant Prompt Caching: Restructured context windows for the YouTube Assistant to maximize prompt caching efficiency, reducing latency and resource usage.
  • Custom YouTube TTS Playback: Optimized playback and queuing logic for custom YouTube TTS workflows.
  • Clear Empty-Audio Error Indicators: Added prominent visual error alerts when importing or processing video files with empty or silent audio streams.
  • Enhanced Font Previews: Fixed an issue where font previews failed to render beside the selection menu. Previews now display comprehensive sample glyphs (Aa, numbers, and Chinese characters) for clear typography comparison.
  • UI Layout Polish: Refined overall layout spacing, alignment, and visual hierarchy across the application.

v1.1.2

August 17, 2026
  • Dedicated Bug Tracking Link: Introduced a new Bug Tracking view to streamline issue reporting, monitoring, and resolution directly within the app.

v1.1.1

August 16, 2026
  • Expanded MCP Voice Options: Added support for Qwen3-TTS and native Apple TTS within MCP Tools, giving you more voiceover choices and automation flexibility.
  • Streamlined macOS Microphone Permissions: Fixed an issue where macOS prompted for recording permissions multiple times—it now asks only once.
  • Read-Along Recording Fix: Resolved an issue where the first subtitle line failed to display in real-time when enabling speech-to-text transcription during script-reading recording.
  • Cleaner Interface: Removed redundant UI elements and clutter to deliver a more focused, distraction-free editing workspace.

v1.1.0

August 16, 2026
  • YouTube Multilingual Subtitles & Auto-Dubbing: Watch YouTube videos with multi-language subtitles and generate instant voiceover dubbing directly in the app—no video download required.
  • Voice Recording with Auto-Subtitles: Record audio directly into your project and let the system automatically transcribe and sync your subtitles.
  • Script Matching (Text-Audio Alignment): Easily align your existing script or transcript with the audio timeline to generate precise subtitles automatically.
  • Expanded TTS & Translation Engines: Added support for Qwen3-TTS, alongside new integrations for Microsoft Edge TTS and Google Translate.
  • Translation Dictionary & Smart Glossary: Maintain custom vocabulary with the new Translation Dictionary, with support for automatically capturing terms and translations directly from your workflow.
  • Structured AI Summaries: Upgraded summary prompts to deliver clean, structured outputs organized into Summary, Key Points, and Conclusion.
  • AI-Refined Interface Localization (i18n): Massively rewritten UI translations powered by LLMs for a more natural and accurate multi-language experience.
  • Fixed known bugs and improved overall app stability for a smoother editing experience.

v1.0.21

August 7, 2026
  • Import SRT Subtitles in Text Editor: You can now directly upload .srt subtitle files into the text editor, making it easier than ever to bring existing text into your project without manual copy-pasting.
  • Quick Refresh for Subtitles & Audio: Updated your transcript or generated a new voiceover? The video editor now supports a single-click reload for subtitles and voice audio, so your video timeline stays instantly synced without re-editing.
  • Fixed known bugs and improved overall app stability for a smoother editing experience.

v1.0.20

August 5, 2026
  • Audio Transcription Queueing: Added a queue system for API audio-to-text tasks, ensuring smoother background processing without system overloads during heavy use.
  • Edge-TTS Voice Support: Added Microsoft Edge-TTS as a new voice option, giving you access to more natural and realistic text-to-speech voices.
  • Video Preview Zoom: You can now zoom in and out on the video preview player to closely check visual details.
  • Localization Updates: Completed missing language translations (i18n) for a seamless multi-language interface experience.
  • Fixed known bugs and optimized overall system stability.

v1.0.19

July 30, 2026
  • Added support for "Skills" and MCP tool integration, allowing the app to connect with external tools and services to expand AI capabilities.
  • Upgraded the underlying AI engine (llama.cpp) to the latest version, improving processing speed and system stability.
  • Added a "Cancel Task" button to all progress bars, giving you full control to stop ongoing tasks at any time.
  • Fixed known bugs and optimized overall system stability.

v1.0.18

  • Added 'Speaker' management tab to easily distinguish and manage different speaking roles.
  • Added a guided 'Workflow Mode' to make the overall operation smoother and more intuitive.
  • Added support for new AI speech recognition models, including Paraformer-zh, Nemotron, and Qwen 3 ASR 1.7b.
  • Added 'Character Count & CPS (Characters Per Second) Segmentation' for more natural subtitle splitting.
  • Added 'Voice Conversion' feature to support changing voice timbre.
  • Added a visual 'Timeline Editor' to easily fine-tune subtitle timing.
  • Fixed an issue where the Qwen 3 alignment function would error out when encountering blank text.
  • Fixed known bugs and optimized overall system stability.

v1.0.17

June 15, 2026
  • Implement API Retry Mechanism: Added an automatic retry mechanism for API requests to enhance connection stability under unstable network conditions.
  • Decouple Core Components: Separated the initialization/activation logic for llama.cpp and whisper.cpp to optimize resource allocation and modular independence.
  • Fix Text Copying Issues: Resolved a bug where certain buttons failed to copy text to the clipboard.
  • Fix Windows Audio Noise: Addressed and fixed the static/noise issue when recording system audio on Windows versions.

v1.0.16

June 11, 2026
  • Video OCR: Added a selection tool to extract hardcoded subtitles from videos.
  • Added support for advanced AI models: VibeVoice, Voxtral, and Distil-Whisper.
  • Scribis for Windows is now available! (.exe setup)
  • Updated the macOS DMG version for better stability.
  • Fixed some issues.