Privacy Policy

Effective Date: July 28, 2026
Last Updated: November 16, 2026 (announced November 9, 2026)

Potential Labs Inc. ("Company," "we," "us," or "our") values your privacy. This Privacy Policy explains how information is collected, used, and protected in the mobile application "Re:spec Alarm" (Android and iOS; "Service"). This policy is intended for users in the United States, Canada, and Australia, and reflects applicable requirements including the California Consumer Privacy Act (CCPA/CPRA), Canada's Personal Information Protection and Electronic Documents Act (PIPEDA), and Australia's Privacy Act 1988.

This policy applies to both the Android app and the iOS app; where processing differs by operating system, this is stated separately in the relevant section.

The Service can be used in a guest state without any account registration or login. We do not collect account-identifying information such as your real name or email address.

We do not sell your personal information for money. The Service displays one banner advertisement at the top of the home screen; no advertising is shown on the alarm ringing screen. The iOS app does not currently display advertising. For users in the European Economic Area (EEA), the United Kingdom, and Switzerland, on both Android and iOS, we obtain advertising consent through Google's User Messaging Platform (UMP) using the IAB TCF framework; if consent is not given, non-personalized ads are shown instead of personalized ads, and if no consent decision is obtained (for example, the consent form fails to load or goes unanswered), no ad is requested. We share a limited set of information — an app-level advertising identifier, device/app information, IP address, approximate location, ad interaction data, and two conversion events — with advertising platforms in order to serve ads and to measure the performance of our advertising, which California law treats as "sharing" for cross-context behavioral advertising. We do not provide advertising platforms with your name, email address, phone number, or anything you have written. California residents may opt out; see Section 9. We can also disable advertising entirely through remote (server-side) configuration.

In this policy, "personal information" means information about an identifiable individual, including information that on its own may not identify you but that could reasonably be combined with other information to identify you.

1. Information We Collect and Purpose

We collect the minimum information necessary to provide the Service, as follows.

Nickname guidance: We recommend using a nickname rather than your real name, and we advise users not to voluntarily enter real-name or other identifying information.

Automatic collection: Usage and diagnostic data may be collected automatically through analytics SDKs during use of the Service. As a mobile application, the Service does not use web cookies. Opt-out methods are described in Section 8.

Advertising identifier: The Service uses the Android Advertising ID (AAID) for analytics, ad serving, and advertising performance measurement. This is a resettable, device-level identifier. You can reset or delete it at any time under Settings → Privacy → Ads on your device.

On iOS, the advertising identifier (IDFA) is used for advertising performance measurement only if you allow tracking through Apple's App Tracking Transparency (ATT) request; if you do not allow it, the IDFA is not collected and advertising performance is measured without the IDFA. After you have responded once in the app, you can change this choice at any time under Settings → Privacy & Security → Tracking on your device. On iOS, we also use the identifier for vendor (IDFV), a device identifier used only among apps from the same developer, for service usage analytics and advertising performance measurement; the IDFV is reset when all of the Company's apps are deleted from the device.

We do not combine these identifiers with information that directly identifies you, such as your name or phone number.

Advertising performance measurement: To measure the performance of our advertising, we transmit, on Android and iOS, app installation and launch records, together with two conversion events (onboarding completed, alarm set), to an advertising platform. On iOS, the IDFA is not transmitted if you do not allow tracking. These transmissions never include anything you have written — custom message text, free-text responses, and speech-recognition transcripts are excluded — and we do not provide the platform with account, email, or phone identifiers. See Section 6 for the recipient.

We process information for ad serving and advertising performance measurement only to the extent described above. The Service displays one banner advertisement at the top of the home screen (the iOS app does not currently display ads); no advertising is shown on the alarm ringing screen. For users in the EEA, the United Kingdom, and Switzerland, on both Android and iOS, we obtain advertising consent through Google's User Messaging Platform (UMP) using the IAB TCF framework; if consent is not given, non-personalized ads are shown instead of personalized ads, and if no consent decision is obtained, no ad is requested. Any marketing-related processing beyond this scope will require separate consent and an update to this policy.

2. Processing of Voice/Audio Information

The Service may process microphone input, speech-recognition result text, and text for voice synthesis in order to provide speech recognition (STT) and voice synthesis (TTS) features.

Speech recognition: The Service processes microphone input in four situations: (1) when you choose to answer by voice during a conversation, (2) when you speak your custom message aloud during an alarm mission, (3) when you give a one-line reply while the alarm is ringing, and (4) when you use Skip by Voice. Speech is converted to text via the operating system's speech-recognition feature (SpeechRecognizer on Android; Apple speech recognition (SFSpeechRecognizer) on iOS). On iOS, speech is recognized only on the device if the device supports it; on devices that do not, it is recognized on Apple's servers. In each case, a screen explaining the purpose of microphone access is shown before the system permission request appears. We do not operate our own speech-recognition server and do not transmit or collect the original audio to our servers. The recognized text is processed as follows:

• (1) Voice answers during a conversation: processed in the same way as answers you enter (Sections 1 and 3).
• (2) Alarm mission: checked on-device against your set custom message and not transmitted to Company servers.
• (3) One-line reply during the alarm: sent, together with your nickname (the name the app calls you), through the Company backend to Google Gemini (Vertex AI) to create a reaction sentence that responds to your reply; the Company backend does not store it.
• (4) Skip by Voice: sent to Google Firebase AI Logic to understand the short phrase you said and decide whether to move past the screen; the Company does not store it.

Voice output (TTS): To read custom messages aloud, the Service either (i) uses the operating system's on-device voice synthesis feature (such as TextToSpeech on Android and AVSpeechSynthesizer on iOS), or (ii) performs voice synthesis via a Company-operated backend (based on Google Cloud Platform) for more natural-sounding audio. In case (ii), the synthesis-target text (custom message) is transmitted over the network to the Company backend and voice-synthesis provider, and the generated audio is temporarily stored in the Company backend's cloud storage (maximum 1 day) before automatic deletion. The text and audio are used solely for voice synthesis, are not used for model training, and related processing/transfers follow Section 6.

Premium voice: If you use a premium voice through a premium subscription or a free promotion, the text to be read (custom messages, fortunes, etc.) is sent through the Company backend to a voice-synthesis provider (ElevenLabs). The Company requests the text and the resulting audio in a mode in which ElevenLabs does not retain them (Zero Retention), and this information is not used to train ElevenLabs models. If premium voice synthesis fails, the default voice (Google Cloud speech synthesis) is used instead.

Depending on how the Android SpeechRecognizer/TextToSpeech, Apple speech recognition on iOS (on devices that do not support on-device recognition), and any recognition/synthesis engine you select operate, audio or text may be transmitted to the servers of the relevant provider (e.g., Google or Apple), in which case that provider's privacy policy applies.

Future self (My voice): If you choose "My voice", the recording of you reading the on-screen consent sentence aloud is stored on your device (excluded from device backup). Each time your future self is created (including previews), the recording is sent through the Company backend to a voice-synthesis provider (ElevenLabs) to create a temporary voice used only for that request, and the temporary voice is deleted right after the audio is made. The text to be read and the resulting audio are requested in a mode in which ElevenLabs does not retain them (Zero Retention), and the recording and temporary voice are not used to train ElevenLabs models. The Company backend does not store or log the recording, and the recording is not used to identify you, such as for speaker verification. If synthesis in your voice fails, the default voice (Google Cloud speech synthesis) is used instead. To animate the speaking character's mouth, the synthesized audio is processed on a Company-operated cloud server (Google Cloud, Republic of Korea region) and is not stored. You can delete the recording at any time in the app's Future self settings; after that, the default voice is used.

You may revoke microphone permission (on iOS, microphone and speech recognition permissions) at any time in your device's app permission settings. Permissions such as "full-screen notification," "battery optimization exemption," and "exact alarm" are used only to display and run alarms on time and do not collect personal information. On iOS, the Service uses Apple's system alarm feature (AlarmKit) to ring alarms, and alarm information such as the alarm time is registered with the device's operating system.

3. AI Custom Message Generation and Sensitive Information

When you select the AI custom-message generation method while adding an alarm, we collect and process your answers to two question flows — (1) an emotion category selection, and (2) a free-text response or conversational reply with the AI — to generate a personalized custom message. Original response text is processed transiently on-device for message generation, and is transmitted to Amplitude for service analytics (Section 6). Questions and answers exchanged in the Chat tab are also transmitted to Amplitude for service analytics.

Responses may contain potentially sensitive information such as your emotions or goals. We do not use such responses for any purpose other than message generation (e.g., model training) without your explicit consent, and this feature is not a substitute for medical or psychological counseling. You may report inappropriate AI responses through the in-app "Report Inappropriate Response" feature or the contact in Section 11. Delegation and cross-border transfer of responses follow Section 6.

For analytics purposes, generated custom messages and the original text of sentences read aloud by the alarm are transmitted to Amplitude (Section 6), and this is used solely to improve the accuracy of message-viewing detection, to check for anomalies in the generation process, and to improve message quality going forward. However, when a generated custom message is output as voice (TTS), the message text may be transmitted to and processed by the Company backend and voice-synthesis provider, as described in Sections 2 and 6.

We analyze your conversation content to infer your emotional state on-device only; this inference result (emotional state, tendencies, needs, etc.) is stored only locally and is not transmitted to or collected by our servers. However, in the course of this inference, the original conversation text is transmitted to Google Gemini for analysis, to which the delegation/cross-border transfer provisions of Section 6 apply.

In Future self, the first step for tomorrow, what happened today, and the name the app calls you (and the previous day's first step and whether you did it) are sent through the Company backend to Google Gemini (Vertex AI) to write your future self's words; the Company backend does not store or log them. Only the character count of these inputs is sent to analytics tools, not the text itself. Because "what happened today" may contain diary-like, potentially sensitive content, we recommend writing only as much as you need. If an input contains an expression indicating a risk of self-harm, no words are generated and information on where to get help is shown instead. Delegation and cross-border transfer follow Section 6.

If you use the fortune theme, the birth date, birth time, gender, and birthplace you enter are sent to the Company backend to calculate today's fortune. The Company backend uses this information only to calculate the fortune and does not store or log it. Your birth date and birth time are not sent to the AI model; only the result calculated from them and your gender are sent to Google Gemini (Vertex AI) to write the fortune text. Delegation and cross-border transfer follow Section 6.

If you use the weather theme, the region you selected from the list is sent to the Company backend to create the weather text, and the Company backend does not store it. Weather information is looked up from weather information providers (the Korea Meteorological Administration and the Norwegian Meteorological Institute (MET Norway)) using representative coordinates for the selected region, and no information that could identify you is passed to them. The weather feature does not use your device's location permission.

If you give a one-line reply while the alarm is ringing, the text converted by speech recognition and your nickname (the name the app calls you) are sent through the Company backend to Google Gemini (Vertex AI) to create a reaction sentence, and the Company backend does not store it. If you use Skip by Voice, the short text converted by speech recognition is sent to Google Firebase AI Logic to decide whether to move past the screen, and the Company does not store it. Delegation and cross-border transfer follow Section 6.

If you enter a follow-up question about your fortune, the question, that day's fortune content, and your nickname are sent through the Company backend to Google Gemini (Vertex AI) to generate the answer; the Company backend does not store or log them (only the character count is recorded). If a question contains an expression indicating a risk of self-harm, no answer is generated and information on where to get help is shown instead. Delegation and cross-border transfer follow Section 6.

This section will be amended and disclosed if data-processing methods change.

4. Retention and Use Period

Personal information is destroyed without delay upon fulfillment of the processing purpose or upon app deletion. Alarm, settings, nickname, and similar data are stored locally on-device and are deleted when the app is uninstalled. Where retention is required by applicable law, we retain the information for the period prescribed by that law.

5. Disclosure to Third Parties

We do not, in principle, disclose your personal information externally. However, it may be disclosed in limited legal exceptions, such as pursuant to law or a lawful request by an investigative authority.

6. Delegation of Processing and Cross-Border Transfer

We delegate the processing of personal information to the following service providers. Because these providers are located outside your country (primarily the United States), such delegation involves a cross-border transfer of data.

Objection to cross-border transfer: You may contact the Privacy Officer in Section 11 (privacy@potential-labs.com) to object to the cross-border transfer of your personal information. If you object, we will exclude your personal information from cross-border transfer. However, features that necessarily require cross-border transfer (AI custom message generation, voice synthesis, Future self, the fortune and weather themes, the one-line reply during the alarm, Skip by Voice, service usage analytics, etc.) may then be restricted.

7. Destruction Procedure and Method

Procedure: Personal information whose processing purpose has been fulfilled is destroyed without delay; where retention is legally required, it is stored separately before destruction.

Method: Electronic files are permanently deleted using technical methods that render them unrecoverable, and on-device local data is deleted when the app is uninstalled. Custom message text and voice audio temporarily stored on the Company backend for TTS are automatically deleted upon completion or within a maximum of 1 day. To request immediate deletion of data transmitted to third-party AI services (Google Gemini) during message generation/inference, or of voice files stored on the server, contact the Privacy Officer in Section 11. The Future self voice recording is deleted from your device when you delete it in the app's Future self settings or uninstall the app, and the temporary voice created at ElevenLabs is deleted right after synthesis; it may remain in ElevenLabs backups for up to 30 days before being irrecoverably deleted. Subscription purchase records are automatically deleted 400 days after the subscription expiry date (or, if there is no expiry date, the last verification date).

8. Automatically Collected Information and Opt-Out

The Service does not use web cookies. Usage and diagnostic data for analytics are collected through analytics SDKs, and you may refuse collection via device settings or by uninstalling the app. During analysis, we do not transmit voice recordings or the speech-recognition text of alarm missions, one-line replies, or Skip by Voice. (Amplitude collects event metadata along with device identifiers and approximate (IP-based) location. Location is only an approximate estimate based on IP; we do not use GPS precise location or location permissions.)

Current product analytics use Amplitude and Google Analytics for Firebase (GA4, Android app only); advertising performance measurement uses the Meta (Facebook) SDK, and ad serving uses Google AdMob. GA4's data retention is 2 months for event data and 14 months for user data. Firebase is also used for AI backend integration and App Check; in the iOS app, it is used only for App Check and AI features, without analytics (GA4).

Opting out of advertising measurement: You can prevent identifier-based measurement by deleting your advertising ID under Settings → Privacy → Ads on an Android device, or by turning off tracking for this app under Settings → Privacy & Security → Tracking on an iOS device (after you have responded once to the in-app request, it can be changed only in this setting). We also maintain the ability to disable advertising performance measurement entirely through server-side configuration; you may request this via the contact in Section 11. Ad serving can likewise be disabled entirely through remote (server-side) configuration.

Managing advertising consent: Advertising consent is requested when you open the app and first arrive at the home screen. If you are in the EEA, the United Kingdom, or Switzerland, you can change or withdraw your advertising consent at any time, on both Android and iOS, from "Manage ad consent" in the app's settings. This item appears only in regions where consent is required and is not shown elsewhere.

9. Your Rights

You may at any time request access to, correction of, deletion of, or restriction of processing of your personal information. On-device local data is deleted immediately when the app is uninstalled, the Future self voice recording can be deleted directly in the app at any time, and permissions such as the microphone can be revoked in device settings. Other requests may be made to the Privacy Officer in Section 11 (in writing, by email, etc.), and we will act without delay. These rights may be exercised through a legal representative or authorized agent.

For California residents (CCPA/CPRA): You have the right to know what personal information is collected, the right to request deletion, and the right to opt out of the sale or sharing of personal information. We do not sell your personal information for money. We do share an advertising identifier, device/app information, IP address, approximate location, ad interaction data, and two conversion events with advertising platforms for ad serving and advertising performance measurement, which California law treats as "sharing" for cross-context behavioral advertising. To opt out, contact us using the details in Section 11 or use the "Do Not Sell or Share My Personal Information" link on this site; we will disable advertising measurement and personalized ad serving for your device. We will not discriminate against you for exercising these rights.

For Canadian residents (PIPEDA) and Australian residents (Privacy Act 1988): You have the right to access and correct your personal information and to lodge a complaint with the relevant regulator (see Section 13).

10. Age Restriction

The Service is intended for users aged 13 and older. Use by children under 13 is not permitted. We do not knowingly collect personal information from children under 13, and if such collection is confirmed, we will destroy it without delay. This is consistent with the U.S. Children's Online Privacy Protection Act (COPPA).

11. Privacy Officer

We designate the following Privacy Officer to oversee personal information processing and to handle user inquiries, complaints, and remedies.

Privacy Officer: Chunghwan Kim (CEO)
Department: Operations Team
Contact: privacy@potential-labs.com / +82-10-7940-5395

12. Security Measures

Administrative: Minimization of personal information processing; restriction of processing authority and access control.

Technical: Data transmitted externally is protected with encrypted communication (HTTPS/TLS), and backend requests are subject to integrity verification (Firebase App Check — Play Integrity on Android, App Attest on iOS). Most user data, such as alarm and settings data, is stored in on-device storage rather than on Company servers. However, custom message text and generated voice audio for TTS are processed/stored on the Company backend for a maximum of 1 day before automatic deletion. The Future self voice recording is stored only in on-device storage and is excluded from device backup. Recordings used for speech recognition are not saved as files.

13. How to Lodge a Complaint

You may contact the following authorities for dispute resolution or consultation regarding privacy:

• United States — Federal Trade Commission (FTC): www.ftc.gov
• Canada — Office of the Privacy Commissioner of Canada (OPC): www.priv.gc.ca
• Australia — Office of the Australian Information Commissioner (OAIC): www.oaic.gov.au

14. Changes to This Privacy Policy

1. This Privacy Policy took effect on July 28, 2026, and the amended policy takes effect on November 16, 2026.

2. (Amendment history) November 16, 2026 (announced November 9, 2026): Added subscription purchase information, such as the purchase token, processed and retained on Company servers to verify subscription payments, and the processing of fortune follow-up questions to the information collected (Sections 1, 3, and 7); disclosed that, when a premium voice is used, the text to be read is sent to ElevenLabs (Sections 1 and 2); and added Google Play purchase verification and Google Cloud (Firestore, Republic of Korea) as processors and broadened the transfer entries for Google Vertex AI Gemini and Eleven Labs Inc. (Section 6). In addition, with the launch of the iOS app, disclosed App Tracking Transparency (ATT) consent, the use of the advertising identifier (IDFA) and device identifier (IDFV), Apple speech recognition, app integrity verification (App Attest), and the application of advertising consent (UMP) and advertising performance measurement on iOS (Preamble and Sections 1, 2, 6, 8, and 12); and added the processing of the fortune theme (birth date, birth time, gender, birthplace), the weather theme (selected region), and the one-line reply during the alarm and Skip by Voice (speech-recognition text), the information the iOS app stores on the device (fortune information, weather region, voice alarm sounds, diagnostic log, and the device identifier in the Keychain), as well as the items transmitted to Amplitude for service analytics (nickname, birth year, questions and answers in the Chat tab, and the original text of sentences read aloud by the alarm) (Sections 1, 2, 3, 6, and 8).

3. (Amendment history) October 29, 2026: With the introduction of Future self, added the voice recording, evening input (tomorrow's first step, what happened today, the name the app calls you), and the generated words and audio, with how each is processed and retained (Section 1); described on-device storage of the recording, its transmission through the Company backend to a voice-synthesis provider, and how to delete it (Sections 2, 7, 9, and 12); and described how evening input is used to write the words and the guidance shown for expressions of self-harm risk (Section 3). Added Eleven Labs Inc. (United States, Singapore, etc.) and Google Vertex AI Gemini via the Company backend (Japan, etc.) as processors and cross-border recipients (Section 6).

4. (Amendment history) September 15, 2026: Disclosed that the items transmitted to Amplitude for service analytics include text written by the user and custom messages delivered by an alarm (Section 1 collection table, Section 3, Section 6), and corrected the previous statement that such content is not transmitted during analysis (Section 8). Speech-recognition transcripts and recordings are still not transmitted.

5. September 10, 2026: Following the introduction of a banner advertisement on the home screen, removed the previous statement that no advertising is displayed inside the app and disclosed that a banner ad is shown on the home screen, that advertising consent is obtained in the EEA, the United Kingdom, and Switzerland (Google UMP / IAB TCF), and that non-personalized ads are shown where consent is not given (Preamble, Section 1). Added Google advertising (AdMob) as a processor and cross-border recipient for ad serving and measurement (Section 6), and described how to change or withdraw consent via "Manage ad consent" in the app's settings (Section 8).

6. (Amendment history) August 11, 2026: Added Meta Platforms, Inc. as a processor and cross-border recipient for advertising performance measurement (Section 6), and specified the scope of the events transmitted and the exclusion of user-authored content (Section 1). Disclosed that the Android Advertising ID is used for analytics and advertising performance measurement, together with how to reset or delete it (Sections 1 and 8). Clarified our position on "sale" and "sharing" under CCPA/CPRA and added an opt-out route (Preamble, Section 9).

7. This policy may be amended in accordance with changes in law, guidelines, or internal Company policy. As the Service has no separate user account (email), changes will be announced through in-app notices and the store listings on app markets (Google Play, App Store). Changes materially affecting user rights will be announced at least 30 days in advance; other changes at least 7 days in advance.

Inquiries regarding this Privacy Policy: privacy@potential-labs.com

Posted URL: https://respec-time.respec.app/en/privacy/

© 2026 Potential Labs Inc.