1. The software comes with a 5th-generation OCR model—a single model supports five script types (Simplified Chinese, Traditional Chinese, Pinyin, English, and Japanese). Overall recognition accuracy has been improved—achieving state-of-the-art accuracy in multiple application scenarios, with a 13 percentage point improvement compared to the previous version, v4! This feature must be specified in the settings. 2. Automatic subtitle position analysis has been added to the edit page. 3. Apply current settings to all videos has been added to the edit page. 4. OCR inference in CPU mode has been changed to f16, resulting in faster inference speed.

Added the ability to upgrade Pro permissions on a monthly subscription type.

Added support for PaddleOCR GPUs, which greatly improves the recognition speed and reduces the usage of the CPU (valid for macOS systems greater than 14.0). Added the ability to switch on and off preview subtitles. By default, the main language of video recognition is set to PaddleOCR, and GPU acceleration is started by default. Fixed bugs.

Completely redesigned! Remove the automatic export SRT subtitle function! Please identify and complete the export using the one click export button. Welcome interface for deleting historical project records! You use the Open Project button to select the project file. Add automatic language recognition function to the speech recognition module. The speech recognition module has added VAD (voice segmentation monitoring) function to reduce the occurrence of auditory hallucinations. The speech recognition module has added support for graphics cards, greatly improving recognition speed. Fixed a bug where the playback factor cannot be used properly. Fixed the bug of inaccurate subtitle associated images. Fixed the memory leak in the PaddleOCR module. Added calculations for the PaddleOCR module using graphics cards. Introduce user account functions and points Added a one click feature to convert traditional Chinese subtitles to simplified Chinese. Added subtitle translation function Added one click translation function (supporting 214 languages), which will consume points Add other features.

1. Add one-click batch export function (SRT and TXT). 2. Add copy setting attribute and paste setting attribute function, which is convenient to set recognition frames, recognize frame intervals, recognize engine and language in batches. 3. Add the function of auto-saving the project, and the project will be automatically saved every time the recognition is completed.

Fixed the bug that UVSOCR in version 2.0 was empty.

Fixed the bug that the time-point frame was misaligned (optical recognition). A new recognition engine, PPOCR (optical recognition runs only on the CPU), has a high recognition rate for special fonts in Chinese Simplified. Fixed some already known issues.

Fixed the bug that the system version is lower than 12.0, and the software will crash when opened

1. Fix the bug that the memory exceeds when extracting large videos, resulting in incomplete recognition. 2. Add the function of copying and pasting the identification frame, which is convenient to quickly set the position of the same identification frame in batches. 3. When proofreading, add the time insertion function at the current time point. 4. When proofreading, add the function of inserting subtitles into the content identified at the current time point. 5. Add the function of adding SRT subtitles. 6. Optimize the bugs that the Acoustic Recognition (Whisper) engine already knows, greatly improving the recognition speed. 7.Fix screen recording video with direction information! Bugs that cause identification errors *Fixed the issue that the recognition rate of optical recognition in version 1.7 caused the recognition rate to drop when calling the laboratory interface

1. Add the function of copying and pasting the identification frame, which is convenient to quickly set the position of the same identification frame in batches. 2. When proofreading, add the time insertion function at the current time point. 3. When proofreading, add the function of inserting subtitles into the content identified at the current time point. 4. Add the function of adding SRT subtitles. 5. Optimize the bugs that the Acoustic Recognition (Whisper) engine already knows, greatly improving the recognition speed.

1. Add audio recognition module, support offline recognition, support Chinese, English, Japanese, Korean and more than 90 kinds of language recognition 2, The screen recognition area is added as the full size of the screen by default 3, Add a button to set the entry and exit points 4, Add a setting panel

1, Add historical item record table 2, Draw the subtitle recognition area frame and add editable function 3, add a new list window, you can add color labels 4, Subtitle editing interface. Add shortcut keybindings: "Space bar": play and pause switch "J" and "K": play back multiplier minus 1 and play forward multiplier plus 1 "Arrow keys left and right": 5 seconds backward and 5 seconds forward 5, Optimize subtitle correction experience

Due to the instability of the new recognition interface of macOS 13.0 SDK. The accuracy of the recognized text is low. This update will be canceled (the new version of macOS 13 will add support languages: yue-Hans->Cantonese Chinese Simplified, yue-Hant->Cantonese Chinese Traditional, ko, KR->Korean, ja-JP->Japanese, ru- RU->Russian, uk-UA->Ukrainian) language support. Wait for it to stabilize before doing compatibility processing.

1. Fix macOS>=13.0 and above, there will be an offset bug when drawing the subtitle recognition area. 2. Repair the generation of srt subtitles, the millisecond time is close to 999.9, which is changed to 000. The bug that caused the time at the exit point to be smaller than the time at the entry point. 3. In the srt subtitle editing interface, add the function of clicking the subtitle block association button to jump to the media playback time point. The new version macOS 13 will add support languages: yue-Hans->Cantonese Chinese Simplified yue-Hant->Cantonese Chinese Traditional ko-KR->Korean ja-JP->Japanese ru-RU -> Russian uk-UA -> Ukrainian

Fix the bug that the system version below the system version 12.0 cannot select the recognition language.

When calibrating subtitles, there is no line break bug when manually adding subtitle blocks.