Your recordings are not a product we sell, a dataset we train on, or something we hand to anyone else. Here is exactly what that means.
Privacy, in plain language
We have never sold a recording or a transcript, never handed one to an
advertiser or a data broker, and never used one to train a model. Not ours,
not anyone else’s. That is not a statement about our intentions — it is a
description of how the service is built, and the rest of this page is the detail behind
it.
We store your account details, service usage and entitlements, and the recordings and transcripts you choose to save so Transcripts works across your devices.
A handful of companies touch your data because delivering the service is impossible
without them. Each one acts on our instructions, for that purpose only, and none of
them is permitted to use your recordings for anything of their own, and none of them is
given more of your data than its job needs. Third-party AI providers transcribe the
audio and separate the speakers. Cloudflare R2 stores saved media. Supabase and
PostgreSQL store account and recording-index data. Google may provide sign-in when you
choose it.
Polar handles checkout, subscriptions, tax, and payment operations as merchant of record. Transcripts does not receive or store your full card number.
No advertising trackers are installed on the public site or in the app. We count page visits, sections reached, call-to-action clicks, demo-language choices, and opened questions ourselves. An optional leaving survey records only the reason you select from a fixed list, alongside the page and campaign information. It has no free-text field. These counts use no cookies and store no IP address or text you enter; Do Not Track and Global Privacy Control are honoured.
A transcript is shared only when you create a link; shared links expire and can be revoked.
You can delete recordings from your library. Use Support for account deletion or a data question.
It gets better at your words, not at everyone else’s expense
A transcript is only good if it knows how you speak. Transcripts is built for
people whose language is usually an afterthought — and it already arrives knowing
that the Arabic spoken across a kitchen table is not the Arabic of a news bulletin, and
that Hebrew as it is actually talked is not Hebrew as it is written. That understanding
is engineered into the product, per language, before you ever sign in. It does not come
from reading anyone’s recordings.
On top of that, the app learns from your library: the names you give the
people who speak, the terms you correct, the voices you choose to have it remember. That
is what makes the second transcript better than the first.
All of it stays inside your account. It is never pooled with other
customers, and no one else’s transcript is ever improved using yours. Two people
using Transcripts in the same room, on the same recording, teach it nothing about each
other.
And nothing you record is used to improve the service for anyone but you.
Not the audio, not the transcript, not your corrections. If we ever want that to change,
we will ask you inside the app, it will be off until you turn it on, and it will be one
switch you can turn back off. Consent that is assumed is not consent, and a policy that
quietly grants it is not a policy we want to have written.
Keeping, deleting and exporting your data
Recordings and transcripts are kept until you delete them. There is no automatic expiry:
nothing is removed behind your back, and nothing is kept once you remove it.
Delete a recording — open it in your library and delete it. The audio
and the transcript go together, from storage as well as from the index.
Export a transcript — copy it, or download it, from the reader. On any
plan, at any time, including after you cancel.
Delete your account — ask through Support and
we remove the account with its recordings and transcripts.
Where processing happens. Transcription runs on third-party AI providers
in the United States, so using the service means your audio is processed there. Your account and
recording index are held in the European Union (Frankfurt, Germany), and the app itself
runs in the European Union (Belgium). Saved media sits in Cloudflare R2, in Cloudflare's
Eastern Europe region — that is where Cloudflare places it rather than a guarantee
it can never be served from elsewhere.
Cookies. We set only what signing in requires — a session cookie that
keeps you logged in. There is no advertising or analytics cookie, and the visit and engagement counting
described above uses no cookie at all.
Who to ask, and who this is for
Contact. Use the form on Support for anything
about your account, a recording, privacy, a refund or billing, and include the account
email we should reply to. If you would rather write directly, the address is
support@transcriptsapp.com. Never send
card, bank or identity documents to either — Polar is the only place for those.
Security reports have their own contact, at
/.well-known/security.txt.
Recording other people. Whether you may record a conversation is decided
by the law where you and the other people are, not by this service. You are responsible
for having whatever consent that law requires before you upload a recording.
What it costs and how to leave is a separate page on purpose, so this one
can stay about your data: see Terms for the plan, the free trial,
cancellation and refunds. Transcripts is not intended for children under 16.