Privacy policy

Podcapp · last updated 1 September 2026 · beta

Podcapp keeps what you send it — links, text, forwarded newsletters — to turn it into your daily audio briefing. That content is passed to AI providers that extract it, analyse it, check it and read it aloud. It is not sold, not used for advertising, and not used to profile you.

Who is responsible

Louis Guichard, an individual, in France. Podcapp is a personal project in closed beta. Any question or request: guich.studio@gmail.com.

What Podcapp keeps

• Your email address — it identifies your account, and it is what recognises the newsletters you forward.
• Two tokens — one for the app, one for your private RSS feed.
• What you capture — the page address, the extracted text, the title, author, publisher and publication date, plus the analysis and the vector derived from them.
• Your episodes — the outline, the script, the sentence-by-sentence verification report, what it cost to make, and the audio file.
• Your settings — language, voice, target length.
• A technical log — a rejected inbound email, a failed extraction: enough to understand why something did not work.

The iOS app carries no analytics, no advertising tracker and no third-party kit. It reads neither your contacts, nor your location, nor your browsing history: it only ever sees what you explicitly share with it.

What leaves, and to whom

Making an episode means sending the content of your sources to specialist providers. Each one receives only what it needs:

• Jina AI — receives the shared address and the text of the page, to extract the content and compute similarity vectors.
• DeepSeek — receives the text of the sources, then the sentences of the script, for analysis, grouping into stories, editorial selection and verification.
• Anthropic — receives the retained facts and the draft, to write and edit the script.
• ElevenLabs — receives the final script, to turn it into a voice.
• Postmark — receives the newsletters you forward, and the address you forward them from, to handle inbound mail.

Postmark aside, whose job it is, none of these providers receives your email address or your tokens: they only see the content to be processed.

The infrastructure itself runs on Neon (database), Cloudflare R2 (audio, feed and build artifacts), Vercel (the API) and Trigger.dev (long-running work). These companies are mostly established in the United States, so your data is transferred there under their own contractual commitments.

What is public, and what is not

The rest — your sources, their text, the verification report, your settings — lives in the database, reachable only with your app token.

How long

As long as your account exists. There is no automatic purge today: deletion happens on request, by hand, within a few days.

Your rights

You can ask for access to your data, its correction, its erasure, its portability, or object to its processing, by writing to guich.studio@gmail.com. An answer comes within thirty days. An erasure removes the account, the sources, the stories, the episodes, the audio and the feed: it is final and nothing is kept. If you are in the EU you may also complain to your data protection authority; in France, the CNIL.

Security

Everything travels over HTTPS, and every account has its own tokens. One acknowledged limit of the beta: forwarding newsletters recognises a user by the sending address, which can be spoofed. Messages that fail SPF authentication are refused, and every refusal is logged.

TestFlight

During the beta the app is distributed through TestFlight. Apple collects install and crash information of its own, under its own privacy policy, and may pass us crash reports if you allow it.

Children

Podcapp is not intended for people under sixteen.

Changes

Any change to this text will be published here, with its date. A change to the nature of the data collected will be announced to you by email.