Last updated 21 August 2026.
What Scribe is
Scribe is a private, invite-only service that turns a video or audio link you
submit into a transcript and a summary. Accounts are created by an operator; there
is no public sign-up.
What we store
- Account data — your email address, display name and the
identifier issued by our authentication provider. Passwords are never stored by
Scribe.
- What you submit — the URLs you send, and any media you
upload, together with the resulting transcript, summary, title, tags and
processing metadata (timings, cost, error messages).
- Operational logs — request and job logs used to run and
debug the service.
Who else processes it
Running a job means sending data to a small number of processors:
- Clerk — authentication (sign-in, session tokens, your
email and profile fields).
- Google — only if you choose “Continue with Google”. Scribe
receives your email address, name and profile picture; nothing is written back
to your Google account.
- Transcription and summarisation providers — the extracted
audio is transcribed on a rented GPU instance, and the transcript text is sent
to a large-language-model provider to produce the summary. Which provider is
used depends on the operator's configuration at the time of the job.
- Object storage — uploaded media is kept in a private
bucket when the upload feature is used.
Scribe does not sell your data, does not serve advertising, and does not use
your transcripts to train models.
Retention and deletion
Transcripts and summaries are kept until you or the operator delete them.
Intermediate audio is temporary and removed after a job completes. To have your
account and its content deleted, write to the contact address below.
Sharing
Content is private to your account unless you create a share link, which makes
that single transcript or summary readable by anyone holding the link until you
revoke it.
Contact
Questions, access and deletion requests: [email protected].