How to make SRT subtitles locally

The short answer

Transcribe the video with a local speech-recognition tool that outputs timestamps, save the result as an SRT file, proofread it, and load it into your player, editor or upload page. If a platform wants WebVTT instead, convert it with a simple tool such as our free in-browser subtitle converter.

What's inside an SRT file

SRT is a plain-text format. Each caption is a numbered block: a sequence number, a time range, one or two lines of text, then a blank line. Times use hours, minutes, seconds and milliseconds with a comma, like 00:01:02,500 --> 00:01:05,000.

WebVTT (.vtt) is very similar but starts with a WEBVTT header and uses a period for milliseconds. It's the format browsers use for HTML5 video. See SRT vs VTT for the differences in detail.

Local ways to make one

• Command line. Open-source Whisper tools such as whisper.cpp can write SRT or VTT directly. Free and offline, but you manage models and file conversion yourself.
• A desktop app. Apps that transcribe locally and export subtitles save the setup work. Check that the export keeps accurate timing and lets you correct names.
• Type them yourself. For a one-minute clip, a text editor and a player that shows timecode is still a perfectly good method.

How to do it in MediaFind

MediaFind transcribes on your computer with word-level timestamps and exports subtitles directly.

1. Add the folder with your video. Transcription runs locally after the one-time model download.
2. Add any names or terms the speakers use, so they're spelled right in the captions.
3. Open the file's transcript and check it against the video.
4. Export the transcript as SRT (or VTT). The file is written on your machine.
5. Load it into your player, editor or upload page. To burn captions into the picture instead, use the Create timeline editor and render to MP4 or ProRes.

Fixing common subtitle problems

Most issues come from line length and timing rather than wrong words:

• Keep captions short, generally one or two lines, so viewers can read them before they disappear.
• If captions drift out of sync, check that the video wasn't trimmed after the subtitles were made.
• If a player ignores the file, check the extension matches the format, and that the file is saved as UTF-8 so accented characters display correctly.
• Proofread proper nouns and numbers; they're where automatic transcripts go wrong most often.

When local makes the most sense

Captioning services are fine for public videos, but unreleased work, client footage and private recordings are a different matter. Making subtitles locally keeps the video on your machine, and once your library is transcribed you also get search across everything for free. For the transcription side on a Mac, see transcribing video offline, or the standalone MediaTranscribe app for local transcripts and subtitles.

Frequently asked questions

What's the difference between SRT and VTT?

Both are timed text. VTT starts with a WEBVTT header and uses a period before the milliseconds, while SRT numbers each caption and uses a comma. Browsers use VTT for HTML5 video.

Can I make subtitles without uploading my video?

Yes. Local transcription tools, including MediaFind, generate timed transcripts on your own computer and export them as SRT or VTT.

How do I convert SRT to VTT?

Add the WEBVTT header and change the comma in each timestamp to a period, or use a converter such as the free MediaFind subtitle converter, which runs in your browser.

Can MediaFind burn subtitles into the video?

Yes. The Create timeline editor can burn in captions and render the result as MP4 or ProRes.