Yes. To draft your letter, your chat and draft are processed by U.S.-based third-party AI providers, reached through an AI gateway. Those providers do not use your inputs or outputs to train AI models, and we pin zero-data-retention routing — so the underlying provider doesn't keep your prompts or the model's output.
Our own copy of the conversation and draft is encrypted and retained for safety and abuse prevention, with access limited — see "What happens to my letter and my data after I send?" for how long.
If you upload an image for context, it's sent to our AI provider only to extract its text; the file itself is not stored.
What happens to my voice when I dictate a letter?
Dictation never reaches us as audio. If you dictate instead of typing, the speech-to-text conversion happens on your device or in your browser's built-in speech feature — not on PostalGenie's servers.
When you dictate, we do not record, collect, store, or transmit your voice. Only the resulting text you choose to submit reaches us, and it's treated like any other text you type.
The voice note add-on is a different thing, and it does send us a recording — that's what it's for. See "What is a voice note, and what happens to the recording?"
One note: your browser, device, or operating system may process the audio under its own privacy policy, which we don't control. If you'd rather avoid that entirely, just type instead.