Subtitle Converter

Convert existing caption files

Select SRT or VTT files and choose the output format. Text encoding is detected separately for each file. One converted file downloads directly as SRT or VTT; multiple files download together as a ZIP. Every output has a text preview and encoding report. Review any uncertain guesses before using the download. Advanced options provides a manual override if text looks wrong. No video or audio upload is needed.

SRT and VTT timing differences

SubRip files commonly use numbered cues and timestamps such as 00:00:01,250. WebVTT begins with a WEBVTT header and uses a decimal point, as in 00:00:01.250. VTT can also omit the hour component. This converter writes explicit hours, minutes, seconds, and milliseconds in both formats and numbers output cues sequentially. The text's timing stays at the same millisecond values.

Worked example: a short caption

An SRT cue starting at 00:00:01,250 and ending at 00:00:03,500 becomes 00:00:01.250 --> 00:00:03.500 in VTT. A WEBVTT header is added above the cues. Two lines of caption text remain two lines. Convert the VTT back to SRT to restore comma timestamps. Cue identifiers are regenerated, so custom identifiers do not survive a round trip.

Plain captions and formatting loss

This tool preserves caption wording and timing rather than reproducing every styling feature. It removes markup tags, cue positioning, VTT notes, style blocks, region definitions, and extra header metadata. The summary reports files where such material was omitted. Common named text entities are decoded, with reserved characters escaped again for VTT output. Review captions containing unusual entities or embedded markup in the destination player.

Validation and optional timing repair

Validation is the default. Every cue needs recognized timing and caption text, separated from the next cue by a blank line. The end must follow the start. Repair mode normalizes overflowing minutes or seconds and swaps reversed start/end times. For example, 00:61.000 becomes 00:01:01.000, and a 5-to-3-second cue becomes 3-to-5 seconds. The report shows before/after timestamps for up to 100 repaired cues per file, plus the total repair count. Missing, malformed, or equal times still stop conversion. Check repaired captions against the recording: these changes cannot establish the intended synchronization. Overlapping cues remain allowed; cue order is preserved.

Automatic text encoding detection

Auto checks Unicode byte-order marks and recognizable UTF-16 byte patterns, then validates UTF-8. For legacy files it analyzes caption text to estimate an encoding, including Western European, Cyrillic, Arabic, Japanese, Chinese, and Korean formats. A batch can contain different encodings. The report identifies the encoding used for each file and distinguishes detected Unicode, estimates, and manual overrides. Short samples or competing interpretations receive an uncertain-encoding warning. Detection is an estimate, not proof of the original encoding: inspect accented and non-Latin characters even when no warning appears.

If text looks wrong, expand Advanced options, select the original input encoding, and convert again. A manual choice applies to all selected files, so convert files separately when they need different overrides. Damaged Unicode files carrying a byte-order mark are rejected rather than reinterpreted as legacy text. UTF-32 is unsupported. Converted output is always UTF-8.

Privacy and limitations

Choose up to 20 files totaling 10 MB, with at most 50,000 cues in each file. Batch ZIP entries have numbered filename prefixes to prevent collisions. Each file's preview shows up to 20,000 characters; downloads contain the complete conversion. Legacy detection examines a sample of up to 128 KB, then the chosen decoder reads the full file. Processing takes place locally in your browser. Keep original files when preserving styling or reviewing timing repairs.