# Lina Edit

> Edit video by its transcript.

Lina Edit turns speech into sentence and word captions on your computer. Select what to remove to cut the matching video and audio, then refine timing, trim pauses and export MP4, SRT or VTT.

- Developer: CommonSource
- Platforms: Windows, macOS 12.0+
- Availability: Official release · Proprietary
- Last updated: 2026-09-30
- Official product page: https://commonsourcelab.com/en/desktop/lina-edit

## Find scenes by what was said

Run speech recognition and every sentence and word stays linked to the video and waveform. Video cuts and caption timing have their own modes, so you can keep a shot and change only when its caption appears.

01. **Transcribe**: A Whisper model splits speech into sentences and words on your computer.
02. **Cut**: Select unwanted words to remove the matching video and audio.
03. **Export**: Tidy captions and pauses, then export MP4, SRT or VTT.

## Key features

### Cutting: Select sentences or words to cut the video

Every transcribed word has a menu to play, select, cut or split. When a boundary is unclear, zoom into the waveform and drag the handles to a frame-accurate range.

- Word menu: play from here, select, cut (X), split before or after
- Adjust caption timing mode, 6.46 to 10.07 s range, Use range as caption time
- Dark theme in Video cut mode, 32.52 to 37.62 s marked for Cut selection

### Refinement: Review silences and repeated words before removing them

Pauses and fillers like “um” and “uh” appear as candidates; listen before and after each cut and apply only the ones you choose. Replace wording in the captions you pick, and mark chapters where the topic changes.

- Normal preset, -38 dB threshold, 0.94 s candidate, Listen before/after cut
- Replacing “wonderful” with “fantastic” in 7 checked results
- 3 chapters and 1 highlight, with Copy chapter list and Keep only highlights

### Captions & export: Style the captions and export the files

Choose 16:9, 9:16 or 1:1 and normalize output loudness. Export a captioned MP4, SRT or VTT subtitles, or only the edited audio as M4A or WAV.

- Pretendard at size 40, caption background, -16 LUFS and English caption options
- Original 320×240, original frame rate, burned-in captions, H.264/AAC

### On-device AI: Install only the speech models you need

Pick a Whisper model by size and accuracy, then download it or import a file. Proofreading and chapter suggestions from a separately installed Qwen model also run locally, and you apply only what you accept.

- Four Whisper models, 74 MB to 547 MB, installed after SHA-256 verification

## Privacy and processing

No cloud processing, sign-in, ads or analytics. The internet is used only for model and demo downloads you start, and your videos, captions and projects stay on your computer.

### Your footage is processed on your computer

Speech recognition, editing and export run on your computer with the models you install. No account or media upload is needed, and the app goes online only when you start a model or demo video download.

## Stores

- [Microsoft Store](https://apps.microsoft.com/detail/9nxzq5msf4mz)
- [Mac App Store](https://apps.apple.com/us/app/lina-edit-ai-text-video-editor/id6815314839)

## FAQ

### Does Lina Edit need an internet connection?

Only to download a speech model or demo video. After you download or import a model in Settings, analysis and editing run offline, with no account or media upload.

### Which languages can it transcribe?

Korean, English, Japanese, Chinese, Spanish, French, German, Vietnamese and Indonesian, or the language can be detected automatically. Whisper can also create English translation captions.

### Which Whisper model should I choose?

Clean recordings transcribe well with the smaller models; choose Small for noisy audio. At least 8 GB of RAM is recommended for Small and larger models, and accuracy varies with recording quality and background sound, so review captions and cuts before export.

### Does editing change my original video?

No. The original media is never modified. Edits are saved automatically, with recovery history, backups of previous saves and relinking if you move the source file.