Convert and edit an existing subtitle file
The Subtitle Converter & Timing Editor opens a UTF-8 SRT or WebVTT file, or pasted subtitle text, in your browser. You can edit caption text and times, shift every cue by a fixed offset, review warnings and download SRT, WebVTT or a plain-text transcript. It is free and requires no account. Start with an existing subtitle file; audio transcription, automatic translation and burned-in video captions are outside this tool's scope.
A cue is one timed caption. Its start and end define when it appears; its text can contain several non-empty lines. English, Arabic, Urdu and other valid Unicode text can be edited without translating it. The text preview follows the text direction automatically, while timestamp fields remain left to right. Check the final appearance in your destination player, since fonts, placement and reading direction are partly player decisions.
SRT, WebVTT or plain text?
| Output | Appropriate use | What this tool writes |
|---|---|---|
| SRT | A player or editor that requests SubRip | Consecutive cue numbers, comma millisecond separators and escaped plain caption text |
| WebVTT | A web player or workflow requesting VTT | A WEBVTT header, dot millisecond separators and escaped plain caption text |
| TXT | Reading, reviewing or repurposing an existing transcript | Caption text in the current cue order, separated by blank lines, without timestamps |
WebVTT includes features beyond ordinary subtitle words and times. A source can contain styles, regions, cue identifiers, voice annotations or placement settings. SRT players can also recognize formatting tags. This editor uses a plain-text model rather than promising to preserve those advanced features. Detected omissions appear in a conversion-change list, and preparing a download is blocked until you acknowledge that list. Keep the original whenever formatting matters.
Worked example: delay three captions by half a second
- Select Load example, then Open subtitles. The example contains three captions in English, Arabic and Urdu.
- Check the first cue. It starts at 00:00:01.000 and ends at 00:00:03.000.
- Leave the shift at 0.500 seconds and select Apply timing shift.
- The first cue now starts at 00:00:01.500 and ends at 00:00:03.500. Its two-second duration is unchanged. The other two cues also move later by 500 milliseconds.
- Choose WebVTT, select Prepare download, inspect the output preview, then download the VTT file.
- Use Undo last change to reverse the shift, or Restore imported cues to return to the opened cue list. Prepare another download after making a change.
A positive offset delays captions. A negative offset advances them. The tool accepts up to three decimal places in seconds, so -1.250 means 1,250 milliseconds earlier. If any resulting time would be negative or above the supported maximum, the entire shift is rejected. No cues are automatically clipped or dropped.
Edit individual captions
Select a cue from the paginated list, then change its start, end or caption text. Each page lists ten cues. You can add a new two-second cue after the latest end time, remove the selected cue, undo the last ten cue changes, or restore the imported list. A new cue has empty text and must be filled in before export.
Time fields accept HH:MM:SS.mmm, with a comma also accepted while editing. Minutes and seconds must each be 00–59, milliseconds use three digits, and the end must be later than the start. Importing SRT expects comma timestamps; importing WebVTT expects dot timestamps and its header. The supported maximum is 99:59:59.999, rather than an unlimited number of hours.
Use ordinary caption text in the editor. Typing markup does not turn it into styled HTML. The preview displays it literally, and subtitle export escapes ampersands and angle brackets. Recognized character references from an imported file are decoded once. Unsupported references remain literal and are included in the conversion-change review. Some SRT players handle character references differently, so inspect the actual target player before publishing.
Understand timing checks
An invalid timestamp, an empty caption, an end before its start, unsupported control characters or a blank line inside an edited caption blocks export. A blank line separates cue blocks in these formats; use a new cue when a separate caption is intended. Malformed source blocks stop the import rather than producing a partial file with silently missing captions.
An overlap is a warning, not automatic proof of a mistake. Two people can speak simultaneously, or captions may intentionally share screen time. The tool preserves those timings and identifies another cue involved in each reported overlap. It does not merge speakers, infer speech boundaries or listen to the video.
Out-of-order starts are also reported. Sort by start time puts cues in chronological order without changing their time values. Equal-start cues retain their relative order. This tool requires chronological starts before SRT or WebVTT export; TXT can retain the current order. Sorting does not repair overlapping intervals.
Review formatting and metadata loss
The plain-text editor flattens recognized bold, italic, underline, font and related subtitle tags. WebVTT voice, language, ruby and inline timestamp tags lose their annotations; ruby annotation words remain in the plain text and may need manual editing. Line-break tags become newlines. Recognized ASS override blocks embedded inside caption text are removed. WebVTT notes, styles, regions, header descriptions, cue identifiers and placement suffixes are omitted from normalized exports.
Read the resulting words and line breaks before accepting these changes. If a speaker name existed only in a voice annotation, add it to the caption manually when it needs to be visible. The Download original source action keeps the imported text, including its original line endings and BOM where present, available separately from the edited export. External X-TIMESTAMP-MAP timing is rejected because it requires a media timeline mapping this editor does not apply.
Privacy, supported sizes and cleanup
One source is processed at a time, with a 2 MiB UTF-8 input limit, 5,000 cues, and 10,000 UTF-16 code units per caption. Edited caption text has a combined safety budget of 2,097,152 UTF-16 code units. Prepared output is limited to 8 MiB; the output preview shows at most 16,000 characters, while the download includes the complete prepared file. Import and export each stop after 30 seconds. Large documents and low-memory devices can fail below these limits.
Processing uses a local module worker. The tool does not upload or autosave source files, filenames, captions or exports. Its interaction area is masked for session recording, and existing consent-aware tool events contain fixed action/result labels rather than private text. Your undo history remains in current-tab memory. Reset clears the source, cues, undo history, settings and prepared result, and temporary workers and download URLs are cleaned up. Downloads already saved on your device remain outside Reset.
Changing source text or input format clears the opened document. Editing cues or output settings clears a prepared download so an old result is not presented as current. LF or CRLF line endings and an optional UTF-8 BOM can be selected for the normalized export. Other encodings, audio, video, ASS files and embedded caption tracks are not supported inputs.
Continue a creator workflow
After exporting TXT, use the Word Counter to review its length or the Text Diff Checker to compare revised wording with a previous transcript. The AI Video Script Generator drafts new production copy from a brief; it has a separate server-based AI workflow and does not extract speech from this subtitle file. Browse the Utilities category for other local text tasks.
Use the Case Converter when you deliberately want to change capitalization in an exported transcript. Caption wording is not automatically case-normalized here.
Technical sources
- W3C WebVTT specification — cue timing, ordering and text-track structure.
- MDN Web Video Text Tracks Format — UTF-8, headers, cue text and metadata features.
- Library of Congress: SubRip Subtitle format — SRT structure and player-dependent text formatting.
Sources checked on 26 October 2026. The tool supports the documented subset above; it is not a full WebVTT conformance checker or a player-compatibility certification.