Privacy

YT Transcriber needs no signup and no account, and keeps no database of transcripts. The server processes captions and text in memory; when it needs Whisper it uses temporary audio and transcript files, which it tries to delete once processing ends. A failed cleanup can leave files behind on the server.

The history keeps up to five complete results in this browser's local storage. You can remove them with “Clear history” on the home page. A session identifier is also kept in sessionStorage, which the browser sends to the API so the usage limits can be applied; it is not an account.

If the target language differs from the captions or the transcript, the text is sent to DeepL to be translated. Whisper runs on the server, not in your browser. The player loads content from YouTube and the fonts are requested from Google Fonts; those services receive connection data and may apply their own cookie and privacy policies.

The API and the infrastructure may produce technical records of requests, such as the IP address, the session identifier and errors. Having no database does not mean there are no logs or third-party processing. Do not send confidential content. If you share a link to a moment, it may carry the full result so someone else can open it without running the transcription again.