Keep an original and check the format
SRT commonly uses numbered cues and times such as 00:00:01,500. WebVTT starts with WEBVTT and uses a dot before milliseconds. Choose a UTF-8 SRT or VTT file or paste its contents. This workspace handles basic cues, not every feature in the WebVTT specification. Rich headers, NOTE/STYLE/REGION blocks, positioning and unsupported markup are rejected instead of discarded.
Convert without moving captions
Choose the desired output format and leave the offset at 0. SRT to VTT adds the WEBVTT header and rewrites timestamps with dots; VTT to SRT writes comma timestamps. Cue identifiers are regenerated as sequential numbers. Text, cue order, line breaks within each cue and supported b/i/u markup are retained. Keep the original if another program needs identifiers or rich presentation features.
Measure a constant offset
Find a clear spoken phrase near the beginning. If its caption appears 1.5 seconds too late, use −1500 milliseconds. If the caption is too early, use +1500. Check a phrase near the end as well. If the mismatch grows over time, a constant shift will not fix it: your video edits or playback speed may differ, and this tool does not perform stretching or frame-rate correction.
Resolve invalid times before export
End must follow start. A negative resulting timestamp blocks the job, even if only the first caption crosses zero. For example, a cue starting at 00:00:01,000 cannot be shifted by −1500. Choose a valid shift or deliberately revise the source elsewhere; no leading caption is clipped or removed on your behalf. Malformed blocks and missing separators also stop the conversion.
Review overlaps rather than automatically deleting them
Two speakers or on-screen text can intentionally overlap. The workspace warns when a cue overlaps an earlier cue or appears out of time order, while preserving the input sequence. Use the paged preview to inspect the relevant cues. The download contains all cues, not only the current preview page. Up to 2,000 cues and 100,000 characters are supported within 400 KB.
Test the actual downloaded captions
Load the exported file alongside the intended video. Check the beginning, a middle section and the ending, especially if captions were shifted. The transcript TXT option removes times but retains text, repeated lines and supported markup; it is not translation or speech recognition. Files never leave this browser, and closing the workspace does not save a recoverable editing history.