01 / TWO JOBS
Reading captions is not the same as writing them.
A transcript can come from two places. It can be read from a caption track that already exists on the video, or it can be written from scratch by speech recognition listening to the audio.
Most of the confusion around uncaptioned video comes from mixing those two up. A page that reads caption tracks cannot invent one for a video that has none, however good it is at its job. A page that runs speech recognition does not need a caption track at all, but it does need the audio, which is why it usually asks for a file.
Before you paste a link, decide which kind of request you are making. That single question saves more time than any tip further down this page.
WatchToText does not use one route everywhere, and each page says which one it uses: a YouTube link reads captions, an upload uses speech recognition, and the TikTok page tries captions first and falls back to speech recognition. This article walks through these routes in the order most people meet them.
02 / A YOUTUBE LINK
A YouTube link only works when captions exist.
The YouTube path reads the caption tracks already attached to the video. It does not listen to the soundtrack. There is no guesswork and no audio processing: the tool asks for the tracks, takes one, and shows the text with its timings.
That has a hard consequence. If the video has no caption track, there is nothing to convert. The page says so plainly rather than returning an empty result, and no amount of retrying changes the answer for that link.
When the video is your own, there is a way out: upload the original file and the page runs speech recognition on it. That is a different route with a different limit, described in the next section. What the link route will not do is take a public YouTube address and transcribe it from audio.
If the request itself is refused, the page also offers another way through: open the video on YouTube, copy the transcript text there, and paste it into the YouTube transcript download page, where it is tidied in your browser and can be saved like any other result.
03 / YOUR OWN FILE
When the video is yours, upload the file.
The mp4 to transcript page takes a file instead of a link, which is the route that works without any caption track. MP3, M4A, WAV, MP4, MOV and WebM are the formats the page names. You do not need to extract the audio first: choose the MP4 as it is, and the browser reads the audio inside the file.
Choose the file.
Use the upload button at the top of that page. A file the browser cannot read comes back with a message to try an MP3, M4A or WAV.
Wait while it works through the audio.
The browser splits the audio into short pieces and sends them for speech recognition, then assembles the words in order.
Copy or download.
The result can be copied, or saved as TXT, SRT or VTT.
The limits are worth knowing before you start, because they are enforced rather than advisory: up to 20 minutes and 200 MB per file. A file that is longer or larger is refused, and the status line says why.
The audio is sent for speech recognition and not stored, and the file you download is built on your own device.
04 / TIKTOK AND INSTAGRAM
Short video links do run speech recognition.
The picture is different for short video. On the TikTok transcript page, a video with a caption track is read from that track, and a video without one is transcribed from its audio instead. You get text either way, and the result tells you which route ran. Speech recognition on that page handles up to 20 minutes per video; a caption track has no length limit there.
Instagram works from the audio throughout. The page does not receive a caption file for the post, so it works from the audio and shows the creator's own caption as a title above the transcript, never mixed into it. A post longer than 20 minutes, or a request that fails, comes back with a message rather than a partial transcript.
So a short video link is a link you can paste, while a YouTube video with no captions is a file you have to bring. Same tool, same output, two different asks. Once you have the text, the next question is how much checking it needs: how to get a more accurate transcript covers that.
05 / WHAT YOU GET
One transcript, three files, one clipboard.
Every route ends in the same result card, and the same four ways to take the text away.
- Copy puts the text on the clipboard and creates no file.
- TXT is the reading copy: the words in order, without timings.
- SRT is the subtitle file: numbered cues, each with a start and an end time.
- VTT is the same timed cues in WebVTT form, for a web player.
If a result has no timings, SRT and VTT stay switched off and the text can still be copied.
06 / WHEN IT FAILS
What each failure is telling you.
The video has no usable caption track
On the YouTube link route, this means the video has nothing to read. If it is your own video, upload the original file instead.
YouTube asks the server to sign in
YouTube sometimes treats a request from a server as automated and refuses it, even for a public video that has captions. Open the video on YouTube, copy the transcript there, and paste it into the download page.
The video will not play
Some videos cannot be played at all from outside YouTube, and the captions cannot be checked. A live replay is one example. There is nothing to fix on your side.
The post will not load
A private, removed or restricted post looks the same from outside. On Instagram, that is the message you will see. On TikTok, the equivalent message says the video could not be loaded right now.
The audio has no recognisable speech
A clip that is only music, or only silence, returns a message instead of invented words. Clips that are mostly talking work best.
07 / QUESTIONS
Can WatchToText transcribe a YouTube link with no captions?
No. The link route reads caption tracks that already exist. For your own video, upload the file and speech recognition will run on the audio.
How large a file can I upload?
Up to 20 minutes and 200 MB. MP3, M4A, WAV, MP4, MOV and WebM are the formats the page names.
Do TikTok and Instagram links need captions?
No. When a TikTok video has no caption track, the audio is transcribed. Instagram works from the audio for every post.
Is the uploaded audio kept?
No. The audio is sent for speech recognition and not stored, and the file you download is built in your browser.
READY WHEN YOU ARE
Bring the file you already have.
No caption track, no problem: the words come from the audio in the file on your desk.
Upload a video and get the text ↗