Subtitle timing checker and converter
Shift basic SRT or plain WebVTT timings and export supported caption text.
Your inputs
Your results
Ready when you areYour results will appear here.
How it works
Inspect basic caption cues, shift each start/end by a signed whole-millisecond offset and download plain SRT or WebVTT. Cue text is retained without interpreting markup.
Read the source reference ↗A few things to know
Supports plain-text cues only; rejects WebVTT metadata, regions, styles and cue settings, and caption tags. Does not validate caption meaning or synchronize to media. Negative shifted times, end-before-start and out-of-order starts fail. 2 MB / 5,000 cues.
When to use this tool
A caption timing shift adds the same signed millisecond offset to every supported cue start and end. It does not adjust speech, translate text or determine synchronization from a video. This tool accepts basic SRT or a plain-text subset of WebVTT, checks ordered starts and positive cue durations, and reports adjacent overlaps for review. Negative shifted times stop conversion rather than being silently clamped. Styling, regions and cue settings are rejected so converting between formats does not quietly discard their meaning. Test the downloaded captions against the actual media before publishing.
- Apply a known millisecond offset after an edit.
- Convert plain caption text between basic SRT and WebVTT.
- Inspect ordered cue timing and adjacent overlap warnings.
How to use subtitle timing checker and converter
Supply a supported caption file
Paste or upload UTF-8 captions and select the input format. SRT needs numeric indexes; WebVTT begins with WEBVTT. Plain text only, without tags or cue settings.
Choose shift and output
Enter a signed whole-millisecond offset within one day and select SRT or WebVTT output. A negative shifted start is rejected.
Review and play the result
Inspect cue timings, text and overlap count. Download the output and check it with the actual video; the tool does not inspect media synchronization.
Shift by three quarters of a second
Inputs
The first SRT cue starts at 00:00:01,250 and ends at 00:00:02,500. Shift +750 ms and export WebVTT.
Result
The first cue becomes 00:00:02.000 to 00:00:03.250. Its text remains Hello, world.
What it means
Both boundaries shift equally, so cue duration is unchanged. This does not prove that the cue matches the video.
Common mistakes to avoid
Treating overlap as always invalid
Adjacent overlapping dialogue can be intentional. The count is a review warning.
Silently dropping styling
The supported subset rejects markup and WebVTT settings instead of claiming a lossless styling conversion.
Clamping negative times without a rule
This tool rejects that result. It does not shorten, delete or shift only part of a cue.
Frequently asked questions
Definitions, edge cases, and practical next steps for this tool. Search all tool FAQs ↗
Which caption subset is supported?
Numeric-indexed basic SRT and plain WebVTT cues with optional identifiers. WebVTT NOTE, STYLE, REGION and cue settings, plus angle-bracket markup, are rejected.
What is the shift unit?
Whole milliseconds. +750 adds three quarters of a second; −750 subtracts it from each cue start and end.
What happens if a start becomes negative?
The conversion stops with a shift error. No automatic clamping or cue deletion is performed. Choose another offset or edit the source explicitly.
Are cue durations preserved?
Yes. The same offset applies to both boundaries. The source must already end after it starts.
Does an overlap fail conversion?
No. Adjacent overlaps are counted for review. Out-of-order starts or nonpositive durations are errors.
Are identifiers preserved?
SRT output regenerates numeric indexes. WebVTT output omits cue identifiers. Plain cue text and supported timing are retained.
Can this find the right offset from a video?
No. Supply a known offset and play the output against the media. There is no video upload, speech analysis or automatic alignment.
Which limits and encodings apply?
One UTF-8 input, up to 2 MB and 5,000 cues. Uploaded text takes priority. Shift magnitude is capped at one day; timestamps require three millisecond digits.
Sources and method
Use the references below to check definitions and platform guidance. The model’s scope and limits are explained above. Read our editorial standards.