How to Transcribe a Video to Text (5 Ways)

To transcribe a video to text, run its audio through speech recognition (YouTube automatic captions, Microsoft Word’s Transcribe, the free open-source Whisper model or an AI captions tool), then proofread the result and export it as plain text, or as an SRT or VTT file if you need timed subtitles. Typing it yourself still works for very short clips, but it takes several times the video’s length.

This guide compares five methods, shows how long each takes, and covers the clean-up that turns a rough machine transcript into something you can publish, quote or caption with.

The short answer

To transcribe a video to text, upload it to a speech-to-text tool such as YouTube automatic captions, Word’s Transcribe (Microsoft 365), the free Whisper model or an AI captions app. Wait for the transcript, fix names, numbers and misheard words while you listen, then download it as TXT for reading or SRT/VTT if you need captions with timing.

Key takeaways

• Decide first whether you need plain text (notes, blog posts, quotes) or timed text (captions). Timed files are SRT or VTT.
• Automatic transcription is fast but never perfect. Budget a proofreading pass of roughly the video’s length.
• Clean audio matters more than the tool: one speaker at a time, a close microphone, no loud music.
• Fix the transcript before you translate it or burn it in, because every mistake is copied forward.

Disclosure: we make QuickVeed and show how to do this in it, along with other ways. Details about other products come from their public pages, checked in October 2026 and linked at the end.

Plain text or timed text: what kind of transcript do you need?

A transcript is just the words. A caption or subtitle file is the words split into short lines, each with a start and end time. Most tools can give you both, but it helps to know which one you are after before you start.

If you are unsure, get the timed version. You can always strip the timing out to get plain text, but you cannot add accurate timing to plain text without running it through a tool again. For the difference between the two subtitle formats, see SRT vs VTT.

5 ways to transcribe a video to text

1. YouTube automatic captions

Upload the video to your channel (unlisted is fine if you don’t want it public). YouTube generates automatic captions with speech recognition in more than 60 languages, including Arabic, Bengali, Hindi and Spanish. Processing time depends on the audio, so they may not appear straight away. Then open YouTube Studio, go to Subtitles, choose the video, select Edit for the language, then Options and Download subtitles. Our guide to getting the transcript of a YouTube video covers this in detail.

2. Microsoft Word Transcribe

In Word for Microsoft 365, go to Home, open the Dictate menu and choose Transcribe, then Upload audio. It accepts MP4 as well as WAV, M4A and MP3, and subscribers can transcribe up to 300 minutes of uploaded audio a month. Keep the pane open until it finishes; Microsoft says it can take about as long as the recording. You get text with timestamps and speaker labels that you can add to the document.

3. Clipchamp autocaptions

Clipchamp, Microsoft’s video editor, can transcribe narration on the timeline: open the captions tab, choose the language and select transcribe media. You can fix words in the transcript and download it as an SRT file. It works one language per video.

4. Whisper on your own computer

Whisper is OpenAI’s open-source speech recognition model. Installed with Python, a command like whisper interview.mp4 --model small writes TXT, SRT and VTT files next to the video. Nothing leaves your computer, and it costs nothing, but larger models need a decent graphics card to run at a useful speed.

5. Typing it by hand

For a 30-second clip, typing is fine. For anything longer, it is slow: expect three to five minutes of work per minute of speech, more with several speakers. Use a player that can slow playback down and pause with a keyboard shortcut.

How to clean up an automatic transcript

Every automatic transcript needs a proofreading pass. Do it with the video playing, ideally at 1.25× speed once you are used to it. The same errors come up again and again:

1. Names and brands. People, companies and product names are the most common mistakes. Search for each one and fix it everywhere.
2. Numbers. Check prices, dates and percentages against what was said. “Fifteen” and “fifty” are easy to confuse.
3. Jargon. Industry terms often come out as similar-sounding common words.
4. Filler words. Decide whether to keep “um”, “you know” and false starts. For captions and quotes, most people remove them.
5. Speakers. In interviews, add who is speaking at each change if your tool didn’t.
6. Music and silence. Delete any text the model invented during music or quiet passages, such as a repeated line.

Keep a short list of names and terms for each show or client and give it to whoever proofreads. If your tool lets you give it a hint or vocabulary list, add those words before transcribing.

How to transcribe a video in QuickVeed

QuickVeed’s Captions & subtitles is built for videos you also want to caption or translate. It transcribes with OpenAI’s whisper-1 and, when whisper writes a language in the wrong script or repeats a line (it happens with Bengali, Hindi and some others), runs the audio again with gpt-4o-transcribe automatically.

1. Open Captions & subtitles and upload the video (up to 3 hours), or pick one from your library. The upload keeps going if you switch to another page.
2. Pick a caption style (11 to choose from, such as Bold yellow, Karaoke fill or Classic subtitles) and a position: bottom, middle or top. Leave “Make the captioned video when it’s ready” ticked if you want the burned-in copy made automatically.
3. Press Transcribe. The app estimates about a minute of processing per 10 minutes of video.
4. On the Edit tab, correct any line, click a time to jump to it in the player, and press Save changes.
5. On the Translate tab, tick up to 5 of 32 languages at a time. Each line keeps its timing.
6. On the Download tab, get an SRT, VTT or TXT file for each language. On the Burn in tab, make a captioned copy per language; it is saved to your library.

The transcript costs 1.5 credits a minute, rounded up. Each burned-in copy is 1 credit a minute (3 minimum, videos up to 30 minutes), and translation is billed by AI use, usually a credit or two per language. New accounts get a 14-day trial with 75 credits, no card needed.

If you only need the text, untick the captioned video option and download the TXT file; you pay only the transcript. One limit to know: the Edit tab changes the words of each line, not its timing. If a line needs to start earlier or later, edit the SRT in a subtitle editor (see how to edit an SRT file).

Tips for a more accurate transcript

• Record well. A clip-on or close microphone beats a camera microphone across the room. Every tool does better with clean audio.
• Turn music down. Background music under speech causes skipped and invented words. If you have the project file, transcribe a version without music.
• Set the language. If the tool asks for the spoken language, tell it rather than letting it guess, especially for short clips or mixed-language speech.
• Split very long files if your tool struggles; some free tools have upload limits or time out on long recordings.
• Transcribe the final cut. If you trim the video afterwards, the timing in an SRT will be off. Our subtitle time shifter fixes a constant offset.

A worked example: a 25-minute interview with two speakers. Automatic transcription takes a few minutes. Proofreading takes about 25 to 35 minutes, most of it on names and crosstalk. Typing it from scratch would take two hours or more. That gap is why nearly everyone starts with a machine transcript now.

Frequently asked questions

How do I transcribe a video to text for free?

Use YouTube automatic captions (upload the video, then download the captions from YouTube Studio), Clipchamp’s autocaptions, or run the open-source Whisper model on your computer. All three are free; Whisper needs some technical setup.

Can I convert a video to text online?

Yes. Online captions and transcription tools take an uploaded video and return a transcript you can edit and download as TXT, SRT or VTT. Most charge by the minute after a free trial.

How accurate is automatic transcription?

With clear audio and one speaker at a time, modern tools get most words right, but names, numbers and jargon still need checking. Noisy audio, music and crosstalk lower accuracy a lot.

How long does it take to transcribe a video?

Automatic tools usually take a fraction of the video’s length. Proofreading takes roughly as long as the video. Typing by hand takes three to five times as long, or more.

Can I get a transcript with timestamps?

Yes. Download an SRT or VTT file, which has a start and end time for every line, or use a tool such as Word Transcribe that adds timestamps to the text.

Try it with your own video

Upload a video and get an editable transcript you can download as TXT, SRT or VTT. Upload a video, read the transcript, and judge its accuracy for yourself. QuickVeed plans start at $18/month, and new accounts get a 14-day trial with 75 credits, no card needed. Sign up or see pricing.

Sources

Competitor details in this post were checked against these public pages in October 2026. Products and prices change, so confirm with each vendor before you decide.

• YouTube Help: Use automatic captioning (supported languages)
• YouTube Help: Edit or remove captions (Download subtitles)
• Microsoft Support: Transcribe your recordings in Word
• Microsoft Support: How to use autocaptions in Clipchamp
• OpenAI Whisper on GitHub (open-source speech recognition)