What the app does
Heredavoz helps heritage and lapsed Spanish speakers practice through a hands-free AI conversation. Depending on the goal you choose, the app turns that practice into a short family lesson or a personal recap.
The adult is the app user. Children do not create accounts and are not expected to use the AI conversation directly.
Information stored on your device
The app stores your profile, learning goal, dialect choice, practice topics, commute reminder, kid first names or nicknames and ages that you choose to enter, practice sessions, transcripts, lessons, streak, and cached lesson audio on your device using Apple storage.
We do not maintain server-side copies of your transcripts, lesson content, audio recordings, or family profile.
Information processed by OpenAI
We use OpenAI's API for realtime voice conversation, diagnostic scoring, lesson or recap generation, and phrase audio. The following information may be sent when you use those features:
- Live practice audio and the resulting conversation transcript.
- Your selected dialect, learning goal, practice topics, and session context.
- Kid first names or nicknames and ages you enter when family personalization is enabled.
- Lesson phrases when you request generated phrase audio.
- An anonymous RevenueCat app user identifier used as an OpenAI safety identifier.
OpenAI states that API inputs and outputs are not used to train its models by default. OpenAI may retain abuse-monitoring logs for up to 30 days unless longer retention is legally required. See OpenAI's API data controls.
Cloudflare Worker and service records
A Cloudflare Worker issues short-lived OpenAI credentials, checks subscription status, enforces usage limits, scores diagnostics, generates lessons, and creates phrase audio. It processes your anonymous RevenueCat identifier and request metadata needed to operate and protect the service.
The service keeps operational records rather than user content:
- Short-lived session receipts and compatibility counters that generally expire within 24–48 hours.
- Admission, recovery, terminal-state, and usage ledgers retained for up to approximately 35 days.
- A diagnostic-used marker retained for approximately 395 days to enforce the one-time free diagnostic.
- Network information such as an IP address may be processed by Cloudflare for security and rate limiting.
The Worker does not intentionally persist transcripts, generated lessons, kid profiles, practice audio, or OpenAI credentials as content records.
Analytics
We use TelemetryDeck for privacy-focused product analytics. Events may record actions such as onboarding completion, paywall display, trial start, lesson completion, or a session-cost estimate. We do not set an email, name, or custom account identifier in TelemetryDeck.
Purchases
Apple StoreKit and RevenueCat process subscriptions, trials, purchases, restore purchases, and entitlement status. RevenueCat creates an anonymous identifier for the app. We do not operate an email-and-password account system.
Children's information
A parent may enter a child's first name or nickname and age to personalize the parent's conversation and family lesson. Please do not enter surnames, school names, addresses, medical details, or other unnecessary information about a child.
Microphone and notifications
Microphone access is used only for voice-practice sessions that you start. Notification access may be used for a commute reminder or a subscription-trial reminder. You can change both permissions in iOS Settings.
Your choices and deletion
In Settings you can revoke AI consent, delete session history and lessons, or delete all locally stored app data. Deleting local data does not cancel an App Store subscription; subscriptions are managed through Apple.
Tracking and advertising
We do not use advertising SDKs and do not track you across other companies' apps or websites.
Changes and contact
We may update this policy as the product changes. The effective date at the top identifies the current version. Questions can be sent to diego@neonsummit.co.