A WebVTT file is not a Word script until the cues collapse

Opening a .vtt in Word feels like it should be enough. Word will open it. You then still have cue numbers, --> timing lines, and NOTE or STYLE blocks sitting in the middle of the dialogue. That is not a script anyone wants to comment on.

WebVTT is a little worse than SubRip for this job because the useful speaker data is often a tag, not a line of text. A cue can carry a <v Speaker> voice span. Consecutive cues from the same speaker should become one paragraph. A raw paste keeps every cue as its own broken sentence.

I convert in the browser, with timestamps as a choice rather than a leftover.

1. Export the captions as .vtt from the tool that made them (YouTube Studio, Zoom's VTT download, a caption editor). UTF-8 is the encoding I want. If the file is a .srt, the same converter reads that too, but the voice-tag case is the VTT one.
2. Drop the file, or paste the cue text. Parsing stays on the machine. The captions are not uploaded to build the document.
3. Decide whether the Word file is a review copy or a reading script. Leave timestamps on for a translator, a QA pass, or anything that has to point back at the video. Turn them off for minutes or a client-facing read. The preview updates before download, so I check the first speaker change and the last cue there.
4. Download the .docx. It opens in Word, Word Online, and LibreOffice. Copy formatted is the alternative when I want the preview on the clipboard instead of a new file.

What flattens: karaoke timing, drawings, and font effects become plain timed text. What should survive: speaker labels from [Name], Name:, and those WebVTT voice tags, merged when the same person continues.

I do this with the subtitle-to-Word converter. One file stays free. A folder of caption files is a separate unlock and is not required for a single WebVTT.