Turn a Video Into Text
You said something good in a clip and now you need it written down. A quote for the caption, the script you improvised and never wrote, a line a customer said on camera. Playing it back and typing is the old way. Turning the video into text is the short way.
That’s how you turn a video into text here: upload the clip, read the words, take a TXT or an SRT.
This is not a subtitling service for a 40-minute talk. It is for the clip you have on your phone right now: under three minutes, one or two people talking, and you want the words on the page.
That is what the video to text page on ReelNab does. You upload the file, the speech comes back as text, and you leave with a TXT or an SRT.
Which files work
The tool takes video and audio: MP4, MOV, WEBM, MP3, M4A and WAV. A phone recording, a screen capture with a voiceover, a voice memo, an exported reel. Up to 3 minutes and up to 100 MB per file.
What matters more than the format is what is on the track. Speech that you can hear clearly, close to the microphone, one voice at a time, comes out well. A talking head, a voice memo, a product walkthrough. Two people over each other in a noisy café, a voice buried under a music bed, a clip filmed from the back of a room: those come out patchy, because the machine hears roughly what you hear.
If the clip is longer than three minutes, cut it first and transcribe the part you need. The whole file is checked before anything starts, so a long one is refused right away rather than halfway through.
The file has to be yours, or one you have the right to use. The page works with a file on your device, not with a link. If the video lives on TikTok or YouTube, save it first, then upload the saved file.
Transcribe video online: the steps
- Open the page. Go to video to text and drop the file into the box, or tap it and choose one. A video shows a small preview; an audio file shows a player. Nothing is sent until you press the button.
- Press Transcribe. The sound is pulled out of the video, then recognised. A one-minute clip is usually under ten seconds; three minutes closer to half a minute. The percentage on the page is real progress, not an animation.
- Read the text. It appears in a box on the same page, in the language the clip was spoken in. The language is detected, there is nothing to set.
- Take it with you. Copy text puts the whole transcript on the clipboard. Download TXT saves it as a plain file. Download SRT saves the same words with timestamps.
The recognition runs on ReelNab’s own server with an open speech model. The audio is not sent to any outside service, and the upload is deleted as soon as the text is ready. The TXT and SRT go about a minute after you download them.
If you need the sound rather than the words, that is a different tool: TikTok to MP3 gives you the audio as a file.
TXT or SRT
Both hold the same words. They are for different jobs.
TXT is the transcript as plain text, one phrase per line. Open it anywhere, paste it into notes, a caption, a document, a translation tool. This is the one you want when you need the words.
SRT is a subtitle file. Every phrase comes with a start and an end time, like 00:00:00,000 --> 00:00:04,200, followed by the line itself. Video editors and most players open it directly. Drop the SRT next to the MP4 in VLC and the captions appear. Import it into CapCut, Premiere or DaVinci and you get captions you can style, instead of typing them in by hand. This is the one you want when the words go back onto the video.
On an iPhone both files land in the Files app, in the same place your other downloads go. If you cannot find them, this guide shows where Safari puts them.
When it gets words wrong
It will, sometimes. This is speech recognition, not a stenographer, and it is a small model on a small server, chosen so the page stays free and fast.
The mistakes follow a pattern, so you can look for them:
- Names and brands. It writes what it hears, so a product name can come back as a similar-sounding common word.
- Numbers. Prices, dates and phone numbers are worth checking against the clip.
- Fast speech and strong accents. Words get merged or dropped.
- Music under the voice. The louder the bed, the more the model guesses.
- Several people at once. Overlapping speech comes out as one muddled line.
- Mixed languages. A clip that switches languages mid-sentence is transcribed as heard, which is not always as meant.
You copy the text and fix it wherever you paste it. For a one-minute clip that is usually two or three corrections. Read it once with the video playing, and it is done.
What the page does not do, on purpose: it does not translate, it does not label who is speaking, it does not summarise, and it does not burn captions into the video. It turns speech into text. The rest happens in your editor.
FAQ
Is MP4 to text free?
Yes. Upload a clip up to 3 minutes and 100 MB, get the text, TXT and SRT. No account, no watermark on the text, no trial counter.
Which languages does it recognise?
Dozens. English is the strongest, and the languages ReelNab is published in, Russian, Spanish, Portuguese, Indonesian, Hindi and Turkish, work well. The language is picked up from the audio; you do not choose it. It does not translate from one language to another.
Can I transcribe a longer video?
Not in one go. Trim the video first, then upload the part you need. A three-minute clip already takes about half a minute of server time, and the page runs one job at a time so it stays fast for everyone.
Does the video leave my device?
The file is uploaded to ReelNab’s server, where the recognition runs. It is not passed to any third-party service, and it is deleted as soon as the transcript is ready. Nothing is kept.
Can I paste a TikTok or YouTube link instead of a file?
No, this page takes files only. Save the video to your device first, then upload it. Only use clips you own or have permission to use.
One upload, a few seconds, and the words you said are text you can paste, search and put back on the video as captions.