Privacy Policy

Alomi · Last updated 4 November 2026

AiappsUtopia publishes Alomi. This policy explains what stays on your iPhone, what the app sends to our servers, what we keep, and who else receives it. The app has no accounts, no ads and no tracking.

On your iPhone

Your phrasebook, level, streak and conversation scores are stored only on your iPhone and are included in your iPhone's iCloud or computer backup. Conversations themselves are not saved after you close them. Deleting the app deletes this data from your iPhone, and you can delete your phrasebook and history in Profile at any time.

Your voice

When you tap the microphone, iOS turns your speech into text on your iPhone whenever it can. Apple's speech service may process the audio when on-device recognition is not available, under Apple's own privacy terms. We never receive or save your voice. Listening only happens between your tap to start and your tap to stop.

Talking with Mia

Mia is an AI. With your consent, the text of what you say, the conversation so far, and any word you tap are sent over an encrypted connection to our server and on to Anthropic's Claude API, so Mia can reply, suggest corrections, explain words and write your summary. Anthropic processes this on our behalf, does not use it to train its models, and may keep it for a limited time for safety monitoring. We do not store what you say. Please don't share sensitive personal information in a conversation.

Mia's voice

Mia's voice is AI-generated, by OpenAI's text-to-speech API. OpenAI receives only the text of the tutor's lines, which are written not to contain personal details, and of the words and phrases you ask to hear. Mia's replies are written to avoid repeating details such as your name, city, job or health. OpenAI never receives your voice or your own words as you said them. OpenAI processes this on our behalf and does not use it to train its models.

Fixed lines, such as each role play's opening line, are cached on our server for up to 30 days so the same line can be replayed without being generated again, then deleted. Dynamic lines, such as Mia's replies, are not stored on our server: the audio is created when it is played and sent straight to your iPhone. Your iPhone also keeps recent audio in its cache.

What we keep

• An install identifier. A random code created when you first open the app and stored in your iPhone's keychain, and it stays there after you delete the app, so free use carries over a reinstall. Our server keeps only a one-way hash of it, with the platform and app version. It is not linked to your name, email or advertising identifier. If you subscribe, it is linked to your subscription's Apple transaction identifier, which does not include your name or email.
• Usage counts. The time of each conversation's first turn, its turn count (kept for a few hours), and the characters of speech generated each day. We use these to apply free and fair-use limits. They do not include what was said.
• AI usage records. For each request sent to an AI provider: the app and feature, your install identifier, your plan, the AI model, the number of tokens or characters processed, how long it took, whether it succeeded and what it cost us. We use these to run the service and understand its costs, and delete them after 90 days. They do not include your words.
• Subscription records. If you subscribe, Apple sends us the transaction identifier, the plan, its renewal and expiry dates, whether it is a free trial, and any refund. We never receive your name, email address or payment details.

Who else receives data

• Apple handles all payments and sends us the subscription records described above, and may process speech recognition as described above.
• Anthropic receives the conversation text and tapped words, as described above.
• OpenAI receives the text of the tutor's lines and of the words and phrases you ask to hear, as described above.
• Superwall shows our subscription screens. It receives device identifiers (a vendor ID and an alias), device details such as model, OS and language, approximate region from IP, and paywall usage events, to measure and improve our paywalls: for example which screen was shown and whether a purchase was made. It also receives the subscription records from Apple. It does not receive your conversations.
• Our hosting provider stores our server and its database.

We do not sell data, share it with advertisers, or use it to track you across apps or websites.

How long we keep it

Install identifiers and usage counts are kept while they are needed to apply limits and to run the service. Each conversation's session, with its turn count, role play or topic and level, expires after 6 hours. The counters of conversations opened without speaking, level checks and safety checks expire after 24 hours. AI usage records are deleted after 90 days. Fixed audio lines are deleted after 30 days, and dynamic audio lines are not kept. Subscription records are kept as long as accounting and tax rules require.

Your choices and rights

You can stop sharing with the AI providers in Profile. The app stays usable without your consent: only the AI features stop (conversations with Mia, word explanations, summaries and her natural voice) until you agree again, and your phrasebook and progress stay available. To ask what we hold or to have it deleted, email support@aiappsutopia.com. Because we do not know who you are, we may ask for details such as your subscription's order ID from your Apple receipt to find your records.

Children

The app is not directed at children under 13, and we do not knowingly collect their data. We follow AI safeguards for users under 18: the app tells you that its tutor is AI, and what you say is screened for safety concerns, with support information shown instead of the tutor's reply when one is found.

Changes

We will update this page if our practices change. The date at the top shows the latest version.

Contact

AiappsUtopia · support@aiappsutopia.com