Transcript to Word

What transcript to Word accepts

A transcript, for this converter, is timed text or speaker-and-clock text. It is not the meeting recording, not the slide deck, and not a summary someone wrote afterward. The parser looks for SubRip cues, a WebVTT document, ASS or SSA dialogue events, or a Zoom-like block: a speaker line, a clock, and the sentence. Paste and upload accept .srt, .vtt, .ass, .ssa, and .txt. The extension is a hint. The shape of the text decides whether cues appear.

If the preview says no cues were found, the paste is the wrong object. Common misses: an AI recap, a list of action items, a chat pane, a YouTube watch-page transcript with no arrow lines, or a .docx opened and copied as a story with the times buried mid-sentence. Delete everything that is not the spoken transcript and try again, or export the caption file the product actually offers.

One file converts without an account. A folder — many caption files, each becoming its own .docx inside one ZIP — is the paid unlock, together with header and footer text. The pricing page describes that unlock. This page does not change the free file or the price. Batch does not merge separate meetings into one manuscript.

Export formats and the steps that match them

Each product names the download differently. The table is the file to grab, the file to ignore, and the page that owns the quirks. The converter above is the same tool those pages open.

Zoom’s portal path is Recordings, then the cloud recording, then Audio transcript. Teams is the meeting recap’s Transcript download. YouTube is Studio, Subtitles, the language row, Download, and then SRT or VTT. Those clicks change. The shape of a good file does not: you can open it and see either --> cues or a speaker line paired with a clock. If you cannot see either, you do not have a transcript this page can read.

How to convert a transcript to a Word document

1. Export from the product that heard the audio. Do not upload the recording to a second transcriber when the transcript is already in the account. Re-transcribe only when transcription was never turned on and nobody saved a caption file.
2. Open the export. Confirm cues or speaker-and-clock blocks. Cut a summary, a chapter list, or a chat off the top so it does not become the first paragraph.
3. Drop the file on the converter. The right pane should turn into paragraphs within a moment. Cue numbers and arrow lines should be gone from that pane.
4. Set Merge same speaker and Show timestamps for the reader you actually have. Details are in the next section. The preview updates before you download.
5. Download .docx. Open it in Word, Word Online, LibreOffice, or Google Docs via Drive. Keep the export you started from. Comments in Word do not flow back to Zoom, Teams, Otter, YouTube, Whisper, or Descript.

Copy formatted is the same preview on the clipboard, for a section inside a document you already opened. HTML download is a standalone page of that preview. PDF is the browser print dialog, Save as PDF, of the same preview. Pick those when you do not need a new Word file. The .docx is the file to hand someone who will edit.

Speakers: what you can keep

A name survives when it is structured. At the start of a SubRip or plain cue, [Alex] or Alex: becomes a bold speaker. A WebVTT <v Alex> voice tag does the same. An ASS or SSA dialogue line uses the Name column. Zoom plain text uses the speaker line above the clock, or a clock and a name on one line in the shapes the parser knows. A name mentioned in the middle of a sentence stays in the sentence. It does not become a new speaker.

Merge joins consecutive cues that share that label when the gap from one cue’s end to the next cue’s start is about 2.5 seconds or less. Meeting apps and caption tools chop a single answer into frame-sized lines. The join puts those lines back into a paragraph and separates them with a space. It does not add punctuation the export skipped. It does not split a paragraph because the topic changed. A long unlabeled file can become a long paragraph. That is readable. It is not yet a script with turns.

If the preview never changes speaker, the file never labeled the turns. Add [Host] and [Guest] in the caption text when you know who spoke, spelled the same way every time, and convert again. If you do not know, turn merge off, download one block per cue, and name people while you listen. Guessing a name is worse than a missing name. Interview files that are already SubRip are written up on interview SRT to Word.

Two languages inside one cue, or two people talking over each other with no labels, should not be merged. Turn Merge same speaker off so each cue stays a block. Overlapping unlabeled speech otherwise lands in one paragraph, and the range stretches across both. The bilingual pages cover pairs and columns. This page assumes one language unless you change that checkbox.

Timecodes: what you can keep

Show timestamps prints a short start–end range for a human reader, such as 00:01–00:06. Hours show in that label only after the cue passes an hour. The SRT arrow line, with its comma and milliseconds, is not copied into the document. The cue index is not copied either. Uncheck the box and the range disappears from the preview and from the .docx. The sentences and the bold names remain.

Leave the box on for a record: QA against the picture, a translator checking a line, minutes that still have to answer “when.” Turn it off for prose: a client recap, a newsletter draft, a script pass about wording. Download both from the same paste if you need both readers. Name the files for those readers. Do not strip clocks out of the only caption file to get the clean copy. SRT to Word with timecode or without is that toggle on its own. SRT to text starts with the box already off, for a paragraph transcript.

Some exports do not contain a real end time. A Zoom paste that is only a speaker, a start clock, and a sentence has no cue end in the source. Prefer the .vtt when the end of the line matters. A WebVTT or SubRip file already has both ends; the range you see is a short form of those clocks, not a new measurement. A clock that disagrees with the picture has to be repaired in a subtitle editor. The Word file will print the clocks it was given.

Layout in the Word file

Each block is a paragraph. The speaker, when there is one, is bold. The time range, when you asked for it, is italic. The words are the cue text after the label is lifted off the front. Setup, in the More menu, sets the typeface, the body size, and the margins before you download. The choices are Calibri, Georgia, Times New Roman, Arial, and Cambria; 10, 11, 12, and 14 point; 0.75, 1, and 1.25 inch margins. There is no stylesheet upload and no custom template beyond header and footer, which are part of the paid unlock.

Caption styling does not become Word styling. <i> and <b> in an SRT remain characters in the paragraph. ASS karaoke and override tags flatten to the spoken words. WebVTT position and color settings are dropped. That is intentional. Word is not a caption player, and a reader does not want alignment coordinates in the minutes.

After download, normal Word work applies: styles, comments, a heading at the top if you want one beyond the filename title the file already carries, a table if you are aligning two languages by hand. None of that writes back to the transcript. If the brief is a Google Doc, upload the .docx, not the raw SRT. SRT to Google Docs is that handoff. SRT to DOCX shows a worked before-and-after on four cues if you want to see the paragraph shape before you spend the free export on a long file.

Where each product’s transcript actually differs

Zoom and Teams can both hand you WebVTT, and they are different downloads from different accounts. A Teams recap may also hand you a finished .docx. Open it. Otter often mixes a spoken transcript with a recap on the same page; only the spoken part has cues. Descript can export the composition as a document or the subtitles as cues. The document is done. The cues are what you drop here, and they already reflect filler-word cuts and Overdub. This site cannot restore a word Descript removed.

YouTube automatic captions are chopped for the player. Merge is usually what you want, after you have a real SRT or VTT from Studio. SubViewer .sbv uses a comma between two times and has no arrow. It will not become paragraphs here. Download again as SRT or VTT. Whisper files add a text problem the clock will not reveal: repeated lines, one word per cue, and a polite ending the speaker never said. Delete those cues, then convert. The links in the table above are the long version of each of those sentences.

A Premiere sidecar and a CapCut caption export are SRT files with editor-specific ways to obtain them. Burn-in is pixels. A project file is not a transcript. If the only thing you have is the video, this page cannot hear it.

What transcript to Word will not do

It will not transcribe. It will not translate. It will not diarize a file that has no speaker labels. It will not shift timing, score a delivery spec, or certify a Netflix QC sheet. It will not import a PDF, a PowerPoint, or an email thread. It will not open the .docx you already downloaded and “clean it up” — there are no cues in that file to find. It will not push Word comments back into the meeting product.

The upload reads UTF-8. A leading byte-order mark is ignored. If names look corrupted, re-save the export as UTF-8 before you drop it. Files over the size limit are refused, and the converter states the limit. Empty files and wrong shapes produce a preview message, not a blank document that looks successful.

Privacy is the same as the rest of the site: the transcript stays in the browser for conversion. Checkout, if you use the paid unlock, is a separate step on the pricing page. The guide index collects the workflow pages if the table above is not the job you have. All guides is that list. What an SRT file is is the format lesson when the transcript you were sent is SubRip and you need to know what each line means.