Privacy Policy
Last updated: November 2026.
Your video stays on your device
When you upload a video or audio file, it is decoded and transcribed inside your browser, on your device. Your file, its audio and the transcript are never uploaded to a server by Vidrion. A transcript you paste is processed in the browser the same way. There is no account and no sign-up, and Vidrion keeps no server storage of your material — if you close the tab without downloading or copying your pack, it is gone.
The speech-model download
The first time you transcribe, your browser downloads the speech model itself (about 43 MB) from Hugging Face's public library and caches it for later use. That is a download to your device so the transcription can happen locally; your media does not travel in the other direction.
YouTube links
If you paste a YouTube link, the link is sent to Vidrion's oEmbed helper, which asks YouTube for that video's public details — title, channel name and thumbnail — and returns them to your browser. That is all a link can provide; no transcript is fetched, because none is available from a link.
Hosting and fonts
This site is hosted on Cloudflare Pages; like any web host, Cloudflare's infrastructure processes standard technical request data (such as IP address and browser type) to deliver pages and protect the service. The fonts (Sora, DM Sans and JetBrains Mono) are loaded from Google Fonts, which receives the technical request needed to serve the font files.
No analytics, no sale of data
This site does not use Google Analytics or any other analytics or advertising tracker. We do not sell personal data — the tool is built so that your material never reaches us in the first place.
If you email us (see the Contact page), we receive your address and message and use them only to reply. We do not sell or share email addresses.
Changes
If this policy changes, the updated version will be published on this page with a new date.