Privacy
Updated November 11, 2026
Your video link and transcript
When you request a transcript, TranscriptVid contacts YouTube or TikTok directly to retrieve accessible caption tracks. YouTube result thumbnails load from YouTube’s image servers. Public TikTok and Instagram video links may be imported through our server when the platform exposes a downloadable video. Media bytes are streamed to your device, not saved on the server; speech recognition then runs in your browser. Private and sign-in-only videos cannot be imported. Do not submit private or sensitive media. TranscriptVid does not send your transcript to Claude; Copy for Claude only writes text to your clipboard.
On-device file transcription
Selected audio and video files are decoded and transcribed in your browser, without uploading them or their transcripts to TranscriptVid. The first use downloads a speech model from Hugging Face and software from jsDelivr. Those hosts receive ordinary download requests, not your audio. Clear browser site data to remove locally cached model files.
Temporary cache
Caption results have a 24-hour cache expiry. Expired records are excluded immediately and removed during subsequent transcript processing. This reduces repeated provider requests. Raw network addresses are not stored in the application database; daily rotating HMAC hashes support request limits.
Analytics
We retain aggregate daily tool-event counts for up to 90 days; cleanup runs during transcript processing. They do not contain transcript text, submitted video URLs, email addresses or account identifiers. If Google Analytics is configured, it loads only after your choice to allow it. Basic hosting and provider request logs may be governed by those services' policies.
Your choices
Use the tool without optional Google Analytics. To change that choice, clear this site's saved browser data. Downloaded files are saved on your device. This test release is operated by the Peel team.