Auto Subtitle Generator
Turn speech in a short audio or video file into timed subtitles, then review them in the editor or download SRT and VTT.
Audio or video file
This version supports files up to 10 minutes and 100 MB that your browser can decode.
The first run downloads an AI model. Keep this tab open while it loads and processes your file.
Processing speed depends on your device. CPU processing can take several minutes; start with a short clip.
Choose a file to get started.
Processing: Automatically selected for this browser
Your audio and video stay on this device. Speech recognition runs in your browser; only the model files are downloaded.
Generated subtitles
Your generated subtitles will appear here.
Related subtitle tools
How to generate subtitles
- Choose a short audio or video file from your device.
- Select a model and spoken language, then generate subtitles. Wait for the initial model download and local recognition.
- Review the result, download SRT or VTT, or open the subtitles in the editor to correct text and timing.
Frequently asked questions
Are my audio or video files uploaded?
No. The selected media is decoded and transcribed in your browser. An internet connection is needed to download the app and recognition model.
Why does the first run take longer?
The browser must download and initialize the model before recognition can begin. Processing speed also depends on your device, and CPU recognition can be slow.
Can I use the generated subtitles immediately?
Treat them as a draft. Background noise, music, accents, and overlapping speakers can cause errors. Review both the text and timing before publishing.
Generate subtitles locally from audio and video
Create a timed subtitle draft without uploading the source media. This browser tool connects local speech recognition to subtitle editing and SRT or WebVTT export.
Audio and video to timed captions
Choose a compatible local file and generate subtitle cues from its spoken audio. Use SRT for common video workflows or WebVTT for browser video players.
Local processing with a model download
Recognition runs on your device after the model loads. Short clips are a practical starting point because browser decoding, available memory, and CPU speed affect processing.
Review captions in the subtitle editor
Open the generated track in the editor to fix wording, split or merge cues, and adjust timing alongside the original audio or video.