How to use
- On the Files tab, drop subtitle files or click to choose them. You can add up to 20 files at once, 20 MB each. Or switch to Paste and type or paste the content directly.
- Each file shows its detected format, encoding and cue count. The format is detected from the content, not the extension, so a wrong extension is fine.
- Pick an Output format (SRT, VTT, SMI or TXT). The start of the first file’s result appears in the preview straight away.
- Optionally enter a time shift in seconds.
-1.5moves every cue 1.5 seconds earlier,2moves them 2 seconds later. - Press Download. One file downloads as is, several come as a ZIP. Output is always UTF-8 without a BOM.
Format comparison
| Format | Typical use | Time notation | Styling |
|---|---|---|---|
| SRT | Most video players, YouTube uploads | 00:00:01,000 (comma) |
<i> <b> <u> |
| VTT (WebVTT) | HTML5 <video>, web players |
00:00:01.000 (dot), hours optional |
Tags, positioning, classes |
| SMI (SAMI) | Older Korean desktop players | Start time only, in milliseconds (Start=1000) |
HTML tags, several language classes |
| ASS / SSA | Anime fansubs, heavily styled subtitles | 0:00:01.50 (centiseconds) |
Fonts, colors, positions |
| SBV | YouTube Studio export | 0:00:01.000,0:00:03.000 |
None |
All five are accepted as input. Output is SRT, VTT, SMI or TXT; ASS and SBV are input only.
How each format is handled
- SRT: cue numbers are ignored and the
-->time line drives parsing. Missing numbers, extra blank lines, CRLF line endings and a missing final blank line are all fine. Coordinates such asX1:100are dropped. - VTT:
NOTE,STYLEandREGIONblocks are skipped, and cue IDs and settings likealign:startare discarded.<v Name>and<c.class>tags are reduced to their text. - SMI: there is no end time, so each cue ends where the next SYNC starts. A SYNC that contains only
means “clear the screen”, and the last cue stays for 3 seconds. If the file has several language classes (KRCC, ENCC and so on), choose one under SMI language; the default is KRCC or the first class. - ASS:
Dialogue:lines are read following the column order of theFormat:line in[Events], even when it differs from the standard. Override tags such as{\i1}are removed and\Nbecomes a line break. - TXT: dialogue only, one line per cue with no timing. A two-line cue is joined with a space, which suits script cleanup.
Example: the SRT time 00:00:01,000 --> 00:00:02,500 becomes 00:00:01.000 --> 00:00:02.500 in VTT, and in SMI it becomes a pair of <SYNC Start=1000> and <SYNC Start=2500> lines. With a shift of -1.5, a cue that started at 1.000 s starts at 0 instead; times cannot go below zero, so they stop at 0.
Why SMI will not play, and garbled Korean text
YouTube and most web players cannot read SMI. It was built for Windows desktop players, so it has to be converted to VTT or SRT first. Old SMI files are also usually saved as EUC-KR (CP949), which looks garbled when opened as UTF-8. This tool checks for a BOM first, reads the file as UTF-8 if that decodes cleanly, and otherwise falls back to EUC-KR. If the list says EUC-KR, that is how the file was read.
Output is always UTF-8. If you need an EUC-KR SMI for an old player, re-save the result with the Text Encoding Converter.
Tips
- If everything is off by the same amount, a time shift here is enough. For subtitles that drift apart gradually or go out of sync midway, use the Subtitle Editor with two-point sync or a range shift.
- To fix a word across the whole file after converting, try Find and Replace; to count characters, use Character Count.
- Time lines that could not be read are reported under the file name. Only those lines are skipped, and the rest is converted.
Your files and text are processed only in your browser and are never sent to a server.