Security & data

Your readers are not tracked

This is the most concrete commitment we can make, so we make it first. The audio player you embed in your articles collects no personal data whatsoever about the people listening:

• No cookies, no advertising trackers, no third-party ad network.
• No IP address stored. An IP is read in memory for the duration of the request, solely to block bots; it is never written to our databases.
• No user agent, and therefore no browser fingerprint.
• No account, no identifier, no listener profile.

We count one play per article, and nothing else. The recorded event is literally empty of anything that could be tied back to a person. This is not an internal policy that could quietly change: it is checked on every run of our test suite, which fails if personal data ever reappears.

One exception, and it stays on the listener's device: the player remembers their playback position in the browser's local storage so they can resume where they left off. That information is never sent to us, and never sent to you either.

Where your data lives

Your articles, your audio files and your account data are hosted in the European Union (Ireland), and our API runs in the EU as well (Dublin). Database access is partitioned per client: the public key used in browsers cannot read any application data, and our storage cannot be enumerated — nobody can list what exists.

Our sub-processors

We would rather name them than write “partners bound by confidentiality commitments”, a phrase that tells you nothing.

The part that doesn't flatter us

Speech synthesis is operated by Cartesia AI, a US company. Your article text is sent to them to produce the audio: that is a transfer outside the European Union. We cannot produce voices of this quality without them.

More importantly: their terms allow them, by default, to use submitted content to improve their own models. So we will not tell you “your content is never used to train an AI” — we are not in a position to guarantee that on our own, and a promise you cannot verify is worthless. If this point is decisive for you, write to us: we will document the exact state of our settings and the guarantees applicable to your account before you sign anything.

What we do guarantee without reservation: we use your content for nothing other than producing your audio. No resale, no commercial analysis, no in-house model trained on your work.

Your audio files

MP3s are served from a public but unlisted address containing two random identifiers, which makes it unguessable in practice, and our storage cannot be browsed. But let us be precise, because the nuance matters: access relies on the secrecy of the address, not on authentication. It is the same model as an unlisted video.

For an article already published on your site, that is coherent: the audio is no more confidential than the article it comes from. If you work under embargo and generate audio before publication, tell us: that case calls for time-limited signed links, which we enable on request. The “download” setting in your client area hides the player button, but that is an interface choice — not protection.

Retention and reversibility

Your content and audio are kept for the duration of the contract. When you leave, you take your files with you and we delete your data on request. No clause lets us keep your work after you go.

A specific question?

If your DPO or technical team needs a data processing agreement (GDPR article 28), details of our security measures, or an answer to something not covered here, write to contact@wedispatch.fr. We answer with verifiable facts, not reassuring vocabulary.

Updated 15 July 2026. See also our privacy policy, our cookie policy and our service status. The French version prevails in the event of any discrepancy.