Product updates

A little better.
With every release.

See what is new in Quran Caption on desktop and mobile. Find improvements, fixes, and the details behind each release.

Page 9 10 releases

Quran Caption

Quran Caption QC-3.5.20

QC-3.5.20

GitHub

Quran Caption v3.5.20


Styles

  • Added Ayah Banners

    A new global style that lets you add a visual container behind ayahs.
    It works similarly to the existing ayah background style, but uses more polished image-based banners with a more Islamic visual style.

  • Fixed vertical alignment styles in the preview

image

AI Video

  • AI Video can now create projects without requiring a background video

  • Removed debug mode from the AI Video workflow

  • Fixed AI Video provider handling
    → it now correctly uses the AI endpoints configured in Settings


AI providers

  • Added quick presets for text AI providers:

    • DeepSeek
    • OpenRouter
    • Groq
    • Gemini
    • OpenAI

This makes it easier to configure AI features without manually entering every provider setting.


Export

  • Added subtitle JSON export

    You can now export all project subtitles as JSON, next to the existing project data export options.


Subtitles Editor

  • Improved segment marking by length

    You can now mark subtitles based on a minimum and maximum word count, making it easier to find subtitles within a specific length range.


Media

  • YouTube downloads no longer show the VBR warning

    Downloaded YouTube media is now automatically converted after download, so the warning is no longer needed.

  • CBR conversion is now asynchronous

    Conversion now shows progress and no longer blocks the interface.

Quran Caption

Quran Caption QC-3.5.14

QC-3.5.14

GitHub

Quran Caption v3.5.14


Community styles

  • Added community styles system

    → You can now:

    • publish your own styles
    • explore styles shared by the community
    • apply them directly to your project

    This makes it much easier to reuse and share visual setups.

image

Export

  • Added automatic batch size mode (default)

    → The software now:

    • dynamically adjusts batch size in real time
    • keeps RAM usage below ~90%
    • maximizes batch length for better performance

    No more manual tuning needed.

  • Fixed an issue where merged subtitles had a transition to blank between them
    → transitions are now seamless

  • Fixed an issue where unwanted “non-empty blank frames” could appear between subtitles

  • Fixed macOS export rendering
    → styles (background, spacing, underline, etc.) now match preview much more closely

  • Fixed vertical alignment not being applied in exported videos

  • Fixed export showing Arabic subtitles even when disabled

  • Fixed rare crashes when translation editions were missing


AI video generation (beta)

  • Added a new AI video generation mode

    → You can write a prompt (example: “a video about respecting parents”)
    → The AI will:

    • select a reciter
    • select a verse range
    • download audio
    • run segmentation
    • trim translations

    → You only need to apply your style and export

    Requires an AI API key configured in settings.


AI segmentation

  • Improved audio alignment reliability (by @Hetchy)

    → Fixes progressive desync caused by differences between:

    • container duration
    • actual decoded audio

    → Subtitles now stay aligned even on long or problematic audio files

  • This applies to:

    • cloud segmentation
    • local models
    • Whisper
    • imported JSON

Subtitles & rendering

  • Added max lines style

    → Ensures subtitles always fit within N lines
    → Automatically reduces font size if needed

  • Fixed Arabic rendering in grouped subtitles without WBW timestamps
    → Arabic now always displays
    → WBW is applied only if all subtitles within the groups include timestamps

  • Fixed Arabic bracket style rendering


Translations

  • Added new “error” status for translations

    → You can now mark translation issues using:

    • missing words detection
    • overlap validation

    → Helps quickly review trimming quality

  • Added Bangla WBW helper translation


UI & workflow

  • Added project search button in the title bar
    → alternative to CTRL+F

  • Fixed export tab UI issue (scrollbar on small screens)


Developer / internal

  • Refactored export system:
    → split large export script (commands.rs) into multiple focused modules

Settings & integrations

  • Added support for DeepSeek AI endpoint
Quran Caption

Quran Caption QC-3.4.97

QC-3.4.97

GitHub

Quran Caption v3.4.97


Quran Support


Search

  • Added project-wide verse search (CTRL+F)
    → opens a custom search bar
    → quickly navigate to a verse or its translation

Export

  • Fixed an issue with long export filenames on Windows
    → now uses shorter temporary filenames
    → prevents failures when projects contain many surahs

UI

  • Improved update modal progress bar contrast
    → better visibility during updates

  • Fixed wizard layout overflow on small screens
    → improves usability in AI segmentation flow

Quran Caption

Quran Caption QC-3.4.96

QC-3.4.96

GitHub

Quran Caption v3.4.96


Export

  • Fixed an issue where export batches could contain incorrect verses
    → ensuring subtitles are now properly aligned with the correct segments

  • Stabilized the export pipeline
    → removed reliance on fragile/edge-case behavior
    → improves overall reliability, especially on long and complex projects

Quran Caption

Quran Caption QC-3.4.95

QC-3.4.95

GitHub

Quran Caption v3.4.95


Export

  • Fixed an issue where batch size was not respected during export when no silence clips were present
    → could lead to memory crashes

  • Fixed an issue where export could fail on long videos when using:

    • transparent background or fade
    • small batch size

    → the merging step could generate a command too large for Windows to execute


Preview & Layout

  • Improved text fitting behavior:
    → vertical alignment is now ignored while fitting text

    This allows combining:

    • max-height
    • vertical alignment

    → ensuring text always fits within the max height while keeping proper positioning


Timeline

  • Added ability to split (unmerge) visual merges directly from the timeline button

    → no need to right-click → unmerge group
    → does not break the entire group
    → faster and more precise workflow


UI

  • Updated donation component to have more options
Quran Caption

Quran Caption QC-3.4.94

QC-3.4.94

GitHub

Quran Caption v3.4.94


AI segmentation

  • Added a new 100% local segmentation model
    → more accurate than the Muaalem-based model
    → currently supports only one surah per audio file

Subtitles

  • Fixed an issue where the verse number was treated as a word in merged subtitles
Quran Caption

Quran Caption QC-3.4.92

QC-3.4.92

GitHub

Quran Caption v3.4.92

This update focuses on improving editing speed, Word-by-Word (WBW) workflow, and AI segmentation flexibility.


Editing workflow

A new quick editing system has been added to reduce context switching.

image

You can now edit subtitles directly from anywhere:

  • Right-click a subtitle in the timeline to:

    • edit Arabic text
    • edit translation
    • edit WBW styling
      → the editor opens directly inside the timeline
  • Double-click on the video preview:

    • Arabic text → edit Arabic
    • Translation text → edit translation

From this overlay, you can also:

  • split a subtitle at a specific word using WBW timestamps

This allows faster iteration without switching tabs.


Subtitle editing

Splitting subtitles is now more precise and WBW-aware:

  • The D shortcut and right-click → split now:

    • preserve Word-by-Word timestamps
    • automatically split at the nearest word from the cursor position
  • Hold CTRL while splitting to:

    • split exactly at the cursor position

This ensures accurate splitting without breaking WBW timing.


Word-by-Word (WBW)

WBW behavior and styling have been improved:

  • Verse number now inherits the highlight style of the last word
    → ensures visual consistency during recitation

  • WBW timestamps are preserved when resizing a subtitle
    → no need to redo word timing after adjusting start/end

  • Highlight color is no longer required to access other WBW styles

  • Added option:

    • "Always Show Verse Number" for WBW mode
  • Fixed WBW timestamps generated by AI:

    • now always start at 0 and end at the subtitle boundary

AI segmentation

The segmentation system has been expanded and made more robust:

  • Added a new 100% local segmentation model based on Muaalem:

    • supports WBW timestamps
    • works offline
    • less accurate than Quranic Universal Aligner (by Hetchy), but usable alternative
  • Added multiple speech-to-text models to choose from

  • Improved fallback logic:

    • if GPU segmentation fails (quota), it automatically falls back to CPU

Translations editor

  • Added a language picker for the WBW helper:

    • English, French, Urdu, Hindi, Dutch, and Pakistani variants
  • Updated UI:

    • more compact layout
    • easier access to filters and pre-selection tools

Timeline and UI

  • WBW words are now displayed inside subtitle clips in the timeline
    → improves readability and editing precision

  • WBW style section is always visible in Arabic styling
    → includes setup guidance if no WBW timestamps are present

  • Escape key now exits fullscreen preview

  • Fixed an issue where dragging with the mouse could select another clip on mouse release


Styles

  • Added customizable verse number color in Text Styling

Export

  • Fixed an issue where batch size was not respected during export, which could cause memory crashes when no silence clips were present in the project

  • Export speed increased by ~40%

  • Fixed issue where wrong subtitles could appear when fade duration = 0

  • Fixed CBR conversion reducing video resolution to 320x240

  • Removed index numbers from VTT subtitle export


Merging

  • Fixed an issue where merging subtitles across mushaf pages
    could display the wrong verse

Codebase

  • Partial refactor of the codebase
  • Added internal documentation for better maintainability