THE CAPTION REPAIR DESK

Better timing.
Cleaner captions.

Tidy an existing subtitle file, shift every cue, and export SRT or WebVTT. A small repair desk for your next video.

01 / SOURCE

Bring your captions

Choose a subtitle file SRT, VTT, or UTF-8 text · up to 2 MiBOr paste subtitle text

Existing SRT or WebVTT required. This tool edits captions; it does not transcribe audio.

02 / ADJUST

Make a clean pass

Output format

Shift all cues

Positive = later. Negative = earlier. Seconds allow up to 3 decimal places.

Trim and collapse extra spaces Line breaks stay intact. Remove caption markup Strip tags and convert common HTML entities to text.

Cue numbers are regenerated. WebVTT notes, layout settings, styles, regions, and identifiers are not preserved.

03 / EXPORT

Your finished captions

LOCAL PROCESSING

Paste a subtitle or try the example to get started.

Cleaned subtitle text

Your cleaned captions will appear here.

A LITTLE CONTEXT HELPS

Clean captions without losing the thread.

Read the practical subtitle guide

Choose a reference moment

Compare one spoken word with its caption. If every cue is equally late, a negative offset brings the whole file forward.

Keep intentional line breaks

Space cleanup trims each line without merging it. An intentional two-line caption remains a two-line caption.

Preview the export

Conversion keeps timing and words, but drops VTT layout settings. Play the exported file with your video before sharing.

Before you press play

What happens when a negative offset crosses zero?

A cue that starts before zero but ends later is clipped to a zero start. A cue that ends at or before zero is omitted. You will see the count in a conversion notice. If every cue would disappear, conversion stops with an error.

Can this fix captions that drift over time?

This tool applies a constant offset to every cue. It cannot correct a mismatch in video speed or frame rate. If the gap grows as the video plays, use an editor that supports time stretching.

Does it keep positioning and styling?

Caption text markup is retained unless you enable its removal. WebVTT positioning settings, notes, regions, style blocks, and cue identifiers are always omitted. Inline WebVTT timing tags used for karaoke-style highlighting are rejected. Conversion is intended for a plain text and timing export.

Are uploaded subtitles stored?

The file picker reads text into your browser. Subtitle text is not sent to our server, saved to your account, or used to train a model. You can use the privacy page to learn about normal hosting metadata.