1. What this tool does.
WatchToText turns a video’s existing captions into text, and can run speech recognition on audio when a platform provides no captions. The routes work differently:
YouTube links
Paste a YouTube watch, youtu.be, Shorts, live or embed link, or an 11-character video ID. We ask YouTube for the caption tracks already attached to that video and show you one of them as text you can copy, or download as TXT, SRT or VTT. We do not play the video or load an embedded player, and for YouTube links we never touch the video’s audio.
TikTok and Instagram links
If the platform has a caption file on the video, we read it the same way. If it does not, we fetch the public video, keep only its audio, and run speech recognition on it to produce the transcript. Instagram provides no caption files, so when a cached transcript is unavailable, Instagram links use this speech-recognition route.
Uploads
If you upload your own file (MP3, M4A, WAV, MP4, MOV or WebM, up to 20 minutes and 200 MB), we run speech recognition on it to produce a transcript.
There is no translation step. Where a video has an English caption track we prefer it; otherwise we take another track the video already carries. On the download page you can choose among the tracks a YouTube video already has. We do not request the platform’s automatic translation.
2. We are not the platforms.
WatchToText is an independent tool. It is not affiliated with, endorsed by, sponsored by, or operated by YouTube, Google, TikTok, Instagram, or any video platform. “YouTube”, “TikTok”, “Instagram” and related marks belong to their respective owners. We use them only to describe what the tool does.
3. Your responsibility for what you submit.
This site is not directed to children under 13, and you should not submit a link or file that contains a child’s personal information.
- Submit a link or file only when you have the rights or permission needed for its retrieval, transcription and intended use, or when applicable law permits those acts. You must also comply with the source platform’s applicable terms.
- Rights in videos, audio and captions remain with their respective rightsholders. Public availability, attribution, or an educational purpose alone does not give permission to copy or republish a work. WatchToText does not grant you rights in third-party content. Check the permission or legal exception that applies to your particular use.
- Do not use the tool to collect transcripts in bulk, to mirror a channel, or to build a dataset you redistribute.
- Do not submit links or files you would not want processed on our server.
4. What we do not do.
We do not:
- sign in to a platform, use an account, or defeat authentication. YouTube sometimes treats requests from our server as automated and declines to share captions, including for public videos that have them. When that happens we may retry the request once through a relay server we operate (hosted on Vultr); that relay makes its outbound request through a third-party ISP proxy (IPRoyal). If YouTube still declines, we tell you and show you how to copy the transcript from YouTube yourself.
- download, re-host or extract audio from a YouTube video we fetched by link. For TikTok and Instagram links without captions, we do fetch the public video’s audio for speech recognition, as described in section 1.
- create accounts, keep a library, or keep a history of your links. There is no account and no per-user history. Public YouTube captions and TikTok/Instagram transcript results, including the basic metadata returned with them, may be cached for up to 24 hours and reused for requests for the same video. Some YouTube responses with no captions, empty captions or an unplayable video may be cached for up to 6 hours. An unplayable result means we could not check the captions, not that the video necessarily has none. These caches are keyed by the platform and video identifier, with language and track type where applicable, not by a user account. Uploaded-file transcripts are not cached. A cached result may still be returned after the source video becomes private or is removed, until the cache expires or access is disabled through this service.
- promise a transcript. If a video has no caption track and no usable speech, there is nothing to convert, and the page says so.
5. Files you upload.
A file you upload is sent to our server and processed only to produce the transcript you asked for: it is handled in memory for that request and is not stored — no copy of the file itself is kept after your transcript is returned, and the rate-limit and daily usage counters follow the retention periods described in our Privacy Policy. It is not used to train any model, it is not shared with third parties except Cloudflare for the speech-recognition step, and it is not sold. The same applies to the audio we fetch from a TikTok or Instagram link. Do not upload files containing personal, confidential or third-party material you have no right to process: speech recognition processes the entire recording, which the browser sends in short pieces.
6. Accuracy.
Captions are not text we write. YouTube’s help pages state that automatic captions are produced by speech recognition and that their quality varies with pronunciation, accent, dialect and background noise; uploaded captions are written by the creator. We keep every caption event in the order the platform returns it, one paragraph per event, without merging overlapping windows, repairing punctuation, labelling speakers or splitting at pauses — so repeated lines stay repeated and short lines stay short. We do not publish an accuracy percentage, because we have not scored any track against a human transcript. Check anything important against the original video before relying on it.
7. The files you download.
The TXT, SRT and VTT files you generate are yours to use, subject to the rights in the underlying video and to the platform’s own terms. SRT and VTT keep each caption event’s original start and end times, including overlapping cues from automatic tracks.
8. Copyright and our content.
The text, structure and design of this site are ours. You are welcome to read, share and link to our pages; please do not copy whole pages or republish them as your own.
Copyright complaints.
If you believe material made available through WatchToText infringes your copyright, use our copyright complaint instructions. Include enough information to identify the work, the material and where you found it. We review reports and may disable access through this service. We cannot remove the source video from another platform.
Repeat infringement is prohibited. In appropriate circumstances, we may restrict access to this service or to identified material in response to repeated infringement. WatchToText has no user accounts. These terms do not represent that WatchToText has registered a designated DMCA agent or qualifies for a statutory safe harbor.
9. Availability.
The tool depends on what the platforms return and on capacity we pay for. It may be unavailable, rate-limited or changed at any time, and features may be withdrawn. Nothing here obliges us to keep the service running.
We may rate-limit or block access, automated or not, to protect the service and the platforms it depends on.
10. No warranty and limitation of liability.
The tool is provided free of charge, “as is”, without warranties of any kind, express or implied, including fitness for a particular purpose. To the fullest extent permitted by law, we are not liable for any loss or damage arising from your use of the tool or from any transcript, including inaccurate text, missed content, or a request that returned nothing. Where liability cannot be excluded, it is limited to the greatest extent the law allows. Nothing in these terms limits rights you have under mandatory consumer law.
11. Changes.
We may update these terms. The date at the top is the version in force. Continuing to use the site after a change means you accept the updated terms. If we change how submitted files are handled, we will say so on this page for at least 30 days before the change.
12. Law and contact.
These terms are governed by the laws of the People’s Republic of China, without regard to conflict-of-law rules. Questions about these terms: [email protected]. Our Privacy Policy explains how data is handled.