Voice of Self

Voice of Self Privacy Policy

Version: 1.5.0
Last updated: 2026-10-04

This Privacy Policy covers Voice of Self across releases. Feature availability and information handling can differ by release; the relevant sections identify those differences. A planned feature described here applies only when it is available in your release. This scope does not remove obligations for older releases. This notice describes information handling; acknowledging it does not grant consent for an optional purpose. Updates do not retroactively authorize processing or change how information was handled before their effective date. Previous versions of this document remain publicly available in its revision history.

This Privacy Policy explains how Voice of Self processes personal information when you use the Voice of Self iOS app, the public website, direct contact channels, and other services that link to this policy.

Voice of Self is provided by Aleksander Jałtuszyk, an individual sole trader established in Poland (“Developer”, “we”, “us”, or “our”).

Contact:

Prepared Kotlin/KMP release. The next release replaces the older Swift app and uses Firebase anonymous authentication, described in section 1.2. It adds the separate Managed AI choice and withdrawal controls described in section 1.5.1. These controls apply only to releases that contain them; this notice does not update an installed older app. At preparation of this version, Voice of Self has not yet been released on Android.

Summary Of Key Points

Table Of Contents

  1. What information do we collect?
  2. How do we process your information?
  3. What legal bases do we rely on to process your personal information?
  4. When and with whom do we share your personal information?
  5. Do we offer artificial intelligence-based products?
  6. How do we handle sign-in and social login features?
  7. Is your information transferred internationally?
  8. How long do we keep your information?
  9. How do we keep your information safe?
  10. What are your privacy rights?
  11. Advertising, tracking, and campaign measurement
  12. Controls for Do Not Track features
  13. Do United States residents have specific privacy rights?
  14. Do we make updates to this policy?
  15. How can you contact us about this policy?
  16. How can you review, update, or delete data?

1. What Information Do We Collect?

1.1 Information stored on your device

Voice of Self stores journal content locally after processing unless you choose to export, back up, share, or directly send it elsewhere. The app may store the following on your device:

When a newer release imports the older local store, migration can retain original database, settings or recording files with a .migrated suffix. Those originals can be plaintext even when the new store uses application encryption. Exported files and system backups also have their own protection and retention. Migration does not itself prove that every old copy was encrypted or erased.

Prepared Kotlin/KMP iOS releases with the protected migration importer: retained legacy SQLite files and sidecars, settings JSON and individual recording originals remain plaintext at the application layer. The importer requests iOS file protection and excludes those originals and temporary recovery copies from backup. It preserves originals and can recover after an interrupted rename. Completion is blocked if any row or expected recording cannot be recovered; imported recordings are verified after encryption. This does not erase older exports or backups already made. Migration does not delete the new encrypted recordings that share the recordings directory.

Depending on your Apple or device settings, local app data may also appear in device backups, Finder/iTunes backups, iCloud backups, or other storage locations you choose when exporting or sharing files.

1.2 Account and sign-in information

Older releases may keep an existing signed-in identity after an update. Newer releases may create a new Firebase identity for a different backend and restore subscriptions through RevenueCat. Restoring a purchase does not transfer or erase every record associated with an older backend identity. In installation-identity releases, authentication/session identifiers are processed without asking for your email or an interactive sign-in; Settings may expose a Support ID so the relevant records can be located. These identifiers are pseudonymous, not anonymous.

The prepared Kotlin/KMP release uses Firebase anonymous authentication. Firebase issues a persistent user identifier and session credentials, stored in the platform’s secure identity storage. The Support ID uses that existing identifier. It can link backend access and usage records with RevenueCat subscription, purchase and customer alias records. Authentication, network and app-attestation metadata can also be personal data. Not asking for an email or a manual sign-in does not make this data anonymous.

In older releases with interactive sign-in, we may process:

1.3 Subscription, access, and backend usage records

If you use account mode or paid features, we may process backend-linked records such as:

Our backend usage and accounting records are intended to track access, limits, cost, fraud prevention, and service operations. They store usage metadata such as token and cost information; they do not store what you said, raw transcript text, raw audio, or AI response content as account records.

1.4 Website and direct contact submissions

If you contact us through website or email channels, we may process:

Direct contact emails may be processed through your email provider and ours. Please avoid sending more personal information than needed when contacting us. Do not send sensitive journal entries, recordings, transcripts, health information, location history, screenshots, or third-party personal information unless that information is needed for the support request.

1.5 Sensitive information

Voice of Self is a self-reflection and journaling tool, not healthcare, therapy, diagnosis, treatment, crisis monitoring, emergency response, or a substitute for a licensed professional. Your journal content may still contain or imply sensitive personal information if you choose to say or write it. This is not because we ask you to create a sensitive profile; it is because an open-ended reflection app lets you talk about your life. This may include information about:

We do not ask you to create a separate sensitive-profile record for our backend. Instead, sensitive information may appear inside the content you create in the app or directly send to us. Managed AI requests process the content needed to answer the request, then the resulting journal content is stored locally by the app. If content appears to involve self-harm, abuse, illegal conduct, or other safety or acceptable-use risks, Voice of Self may block, hide, refuse, or leave some analysis incomplete.

1.5.1 Sensitive information in entries and transcripts

You choose what to write, say, record, transcribe, export, delete, or send. An open-ended reflection app can include sensitive personal topics such as health, mental health, relationships, sexuality, religion, politics, race or ethnic origin, philosophical beliefs, family matters, work issues, legal concerns, or other private subjects.

Managed AI features may process the current entry, current transcript, live transcript excerpts while a recording is in progress, completed transcript text, and selected past entries, transcripts, summaries, relics, milestones, or related local context depending on the feature, account state, app settings, and workflow.

Managed AI is an optional purpose separate from acknowledging the Terms and Privacy Policy. Where explicit consent is required for sensitive journal information, it must cover the submitted content, purpose and recipients described here; accepting documents or granting an OS speech/microphone permission alone is insufficient.

Releases with the Managed AI choice in Settings > Legal: the choice starts off. You may continue to local journals without enabling managed AI. An explicit choice covers submitted entries, transcripts, live excerpts and selected past context sent through Firebase to the described managed providers. You can withdraw in Settings > Legal, including while offline. Withdrawal on that installation stops new requests, retries and queued/resumed AI work and prevents older replies from being applied. A request already received by the backend or provider may finish; withdrawing cannot recall content already sent. A new explicit choice is required to enable managed AI again or after a material consent-copy change. Restoring app settings does not restore an AI opt-in. Local journal access, export and purchase recovery remain available.

Older releases without that control: the newer control is not added to an older installation by this policy. Stop submitting managed-AI requests and contact voiceofselfapp@gmail.com for withdrawal or rights assistance, identifying the older release/account where possible. Deleting entries or avoiding a feature alone does not prove that previously queued or forwarded processing has stopped. Withdrawal does not by itself delete earlier records, cancel a store subscription or remove accrued purchase rights. Any retention after withdrawal needs its own applicable legal basis.

Voice of Self does not use journal content, raw audio, transcripts, live question content, or AI outputs for ads, retargeting, custom audiences, or training our own models. For normal managed AI requests, request content is handled transiently to fulfill the feature and is not retained as readable backend content after fulfillment.

1.6 Payment information

Paid subscriptions are sold through Apple’s in-app purchase system. Apple handles payment card or billing instrument details. We may receive limited subscription and transaction status information from Apple or RevenueCat, such as product identifiers, entitlement status, renewal state, and related timestamps, but we do not receive or store your full payment card number.

The app may use subscription status already available to it to schedule a local device reminder about 24 hours before an account-mode free trial ends or before a paid App Store subscription period ends. This reminder uses the app’s existing notification permission, is generated on your device, and does not add new backend data collection, retention, deletion, or third-party sharing.

1.7 Information processed automatically

When you use the website, sign in, call backend features, or use subscription or AI-connected services, we and our service providers may automatically process limited technical information such as:

If you use the app, it may also request access to features such as:

You can control device permissions in iOS Settings.

1.8 Information from other sources

We do not buy personal data from data brokers, public databases, or marketing partners.

We may receive limited information from third parties that help us deliver the service, including:

1.9 Children

Voice of Self is intended only for adults aged 18 and over. We do not seek to provide the app or managed journal processing to users under 18. An age declaration or store content rating does not verify age or prove that younger people never use the app. If you believe an under-18 user has provided personal information through the app or direct contact channels, contact us. We will investigate eligibility, further managed-AI processing and applicable erasure or other privacy rights. Existing data rights and mandatory consumer remedies remain available.

2. How Do We Process Your Information?

We may process personal information to:

We do not sell personal data. We do not use your registered email address for optional marketing or giveaway campaigns without a separate consent flow.

Safety checks are automated. They may refuse, defer, block, or limit AI output or show safety-oriented language when content appears risky. Voice of Self is not live human monitoring, crisis support, therapy, or emergency response. Human review or disclosure is limited to situations such as support you request, security, abuse prevention, legal compliance, or urgent safety/legal circumstances described in this policy and the Terms.

If you are in the EEA, UK, Switzerland, or another jurisdiction that requires us to identify a legal basis, we generally rely on one or more of the following:

Where sensitive personal information is processed, it is processed because you chose to include it in content you ask the app to process or because you directly sent it to us. We rely on your explicit action, instructions, consent where required, or other grounds permitted by applicable law.

4. When And With Whom Do We Share Your Personal Information?

We may share information with service providers or infrastructure partners when needed to operate the service. Categories include:

We may also disclose personal information when reasonably necessary to:

We do not sell personal data, and we do not share personal data for cross-context behavioral advertising.

5. Do We Offer Artificial Intelligence-Based Products?

Yes. Voice of Self includes AI-assisted features such as:

AI features are for non-authoritative reflection only. They are not advice, guidance, therapy, crisis support, emergency support, safety-critical decision support, or proof that progress is truly present or absent. Additional safety and acceptable-use limits appear in the Terms of Service.

Current managed provider roles are summarized below. Provider availability, routing, model choices, and infrastructure settings may change over time.

ProviderPurposeData handled for the request or serviceRetention and training notesRegion / transfer note
Firebase / GoogleAuthentication, callable backend, App Check, security, account records, usage/accounting metadata, and backend infrastructureAccount identifiers, authentication/session metadata, request metadata, App Check or integrity signals, token/cost/workflow metadata, performance metadata, and limited operational logsBackend records are designed not to store raw journal text, transcript text, raw audio, or AI output as persistent account records. Platform logs may retain limited technical metadata under Google/Firebase settings.May process in regions used by Firebase / Google infrastructure.
Amazon Bedrock / AWSManaged analysis, live-question, and related AI processingTranscript text, live transcript excerpts, selected entry/context needed for the requested workflow, prompts, outputs, and technical request metadata while servicing the requestWe do not use your content to train our own models. Our backend is designed not to retain raw request or response bodies as persistent journal records. Model/region retention and invocation logging are separate provider configuration, as described in section 8.5.May process in AWS regions used for managed AI processing.
GroqOptional managed AI processing only when backend routing is explicitly configured to use GroqTranscript text, live transcript excerpts, selected context needed for the requested workflow, prompts, outputs, and technical request metadata while servicing the requestWe do not use your content to train our own models. Our backend is designed not to retain raw request or response bodies as persistent journal records. Groq reliability/abuse-monitoring exceptions depend on its zero-retention controls, as described in section 8.5.May process in regions used by Groq and its infrastructure providers.
AppleApp Store billing, Sign in with Apple, DeviceCheck / App Attest, local device permissions, and Apple transcription or Apple-assisted features where usedApple account/platform data, entitlement and transaction status, device integrity signals, microphone/speech data where Apple services are used, and device permission stateApple handles platform data under Apple’s own policies. Local app records remain under local app/device behavior unless you export, back up, share, or send them elsewhere.Depends on Apple services, your device settings, and Apple’s infrastructure.

5.1 Managed account-mode AI

If you use account mode, content needed to fulfill the request may pass through our Firebase backend to the managed AI provider selected by our backend configuration for that workflow. Managed AI processing uses Amazon Bedrock / AWS for the default managed AI routes, including account-mode analysis and live-question workflows with OpenAI GPT-OSS model identifiers such as openai.gpt-oss-20b-1:0 and openai.gpt-oss-120b-1:0. Groq may still be used if backend workflow routing is explicitly configured for a Groq model. This may include transcript text and, in some flows, live transcript excerpts while a recording session is still in progress.

For normal managed AI requests, our backend is designed not to keep transcript text, selected context, prompts or AI output as persistent readable journal records after fulfillment. That design does not establish all provider or technical-log retention settings. Section 8.5 explains the distinction. Raw audio is handled locally or by Apple speech/transcription services where those services are used; it is not stored by our backend as a readable account record. We do, however, retain limited security, subscription, diagnostic, performance, and usage/accounting metadata such as token counts, workflow labels, timestamps, processing duration, and derived cost.

Our app and backend code do not add Bedrock prompt cache checkpoints for these managed requests. Zero-retention use also requires verified model and regional retention settings, disabled raw invocation logging and, for Groq routes, the relevant zero-retention controls. An earlier zero-retention commitment is not withdrawn by this clarification; a route must be checked against the applicable commitment before activation.

5.2 On-device and Apple-assisted processing

Depending on the feature and your device capability:

5.3 How to limit or avoid AI processing

Voice of Self does not currently offer a user-selected custom AI provider or bring-your-own API key mode. You can reduce or avoid off-device AI processing only by:

We do not use your personal content to train our own models. Managed AI providers process data under their own API terms, privacy policies, and service settings.

6. How Do We Handle Sign-In And Social Login Features?

Voice of Self currently supports:

If you use Google or Apple sign-in, we receive only the limited account information needed to authenticate and operate your account. We do not currently support Facebook, X, or other general social-media login systems.

Third-party sign-in providers control their own processing of your data under their own policies.

7. Is Your Information Transferred Internationally?

Yes. Depending on the feature you use and the provider involved, your information may be processed in countries other than your own, including the United States, AWS regions used for managed AI processing, regions used by Firebase / Google infrastructure, and other regions where our service providers operate. If you export or share files, transfer and storage locations also depend on the apps and storage destinations you choose.

If you are in the EEA, UK, or Switzerland, these countries may not always have privacy laws identical to those in your jurisdiction. Where applicable, we rely on the safeguards, contractual commitments, or transfer mechanisms offered by our service providers under their terms and data-processing commitments.

8. How Long Do We Keep Your Information?

We keep personal information for as long as reasonably necessary for the purposes described in this policy, unless a longer period is required by law.

8.1 Local data

Local journal data remains on your device until you delete it, remove the app, or otherwise clear it yourself. Copies may also remain in exported .vos files, Files/iCloud locations, email drafts, sent-mail folders, or device backups until you delete those copies too.

8.2 Recordings

Recordings may be retained locally for playback, transcript repair, or re-transcription.

8.3 Backend-linked account data

Backend-linked records remain until you delete your account, ask us to delete applicable data, or we remove the records for fraud, safety, security, legal, or service-operation reasons. These records may include:

When you delete your account, most of these are removed straight away. Two are kept for a 30-day grace period before being removed. See Section 8.3.1.

8.3.1 The 30-day account deletion grace period

Legacy signed-in flow. This subsection describes releases that provide the signed-in deletion and recovery mechanism. It does not describe the newer installation-identity releases or imply that a missing deletion button exists. Earlier commitments for those legacy users are not withdrawn by this scope clarification. Contact support if the flow is missing or fails in your release.

Deleting your account starts a 30-day grace period. During it, we keep exactly two things:

Everything else listed in Section 8.3 is deleted immediately when you complete the delete-account flow, including your profile record, entitlement overrides, extra usage grants, per-request usage-event records, your RevenueCat customer record, and any files stored under your user prefix. Your existing sign-in sessions are ended immediately, so the app signs you out and the account stops working like an active account.

We keep those two things for two reasons.

The first is recovery. If you deleted your account by mistake, or changed your mind, signing back in with the same Apple or Google identity during the 30 days returns you to the same account rather than a stranger’s blank one. This is the behaviour Apple and Google use for their own account deletions.

The second is abuse prevention. Managed AI usage is metered per account per billing period, and each request costs us real money at our AI providers. If deleting an account issued a brand-new account identifier immediately, anyone could spend a period’s AI allowance, delete, sign back in, restore their subscription, and receive a fresh allowance — repeatedly, and at our cost. Keeping the account identifier and the period counters for the grace period is what prevents that. Thirty days is not arbitrary: it is slightly longer than the longest billing period we sell, so waiting the grace period out gains nothing that simply waiting for your billing period to renew would not already give you.

During the grace period, the retained records are used for nothing else. They are not used to contact you, to profile you, to advertise to you, to train models, or for analytics. The retained usage totals are counters, not content: they contain no entries, transcripts, audio, or AI output, because we do not retain that content in the first place (see Section 8.5).

When the 30 days end, an automated job removes the usage totals and then the Firebase Authentication account record. After that, signing in with the same Apple or Google identity creates a genuinely new account.

If you are in a jurisdiction that requires us to identify a legal basis, we rely on legitimate interests for this retention: preventing fraud and abuse of a paid, metered service, keeping that service financially viable, and allowing people to undo an accidental deletion. Thirty days also aligns with the one-month deadline for responding to data subject requests under Article 12(3) GDPR. You can object to this processing, or ask us to complete the removal sooner, using the contact details in Section 15.

8.4 Diagnostic, security, and platform log records

Operational, security, anti-abuse, and deletion-request records may persist for as long as reasonably necessary to detect incidents, enforce limits, document requests, and satisfy legal or compliance obligations. These records are intended to be limited in content and are not intended as a journal archive. They may contain identifiers, timestamps, request metadata, status codes, model or workflow labels, error names, file paths, or other technical context.

Firebase / Google platform logs for backend functions are handled by Cloud Logging according to the project’s configured Google/Firebase retention settings. Under standard Firebase / Cloud Logging defaults, Cloud Functions for Firebase logs are automatically deleted after 30 days unless custom retention or export settings are configured. Our backend code is designed not to log raw journal text, raw audio, or raw AI output.

8.5 Managed-AI content retention

For normal managed AI requests, our backend is designed not to keep transcript text, selected context, prompts or AI output as persistent readable journal records after fulfillment. Provider retention is separate: Amazon Bedrock’s retention can depend on the model, region and account configuration, and raw invocation logging can retain request/response bodies if enabled. Groq’s ordinary inference behavior has reliability and abuse-monitoring exceptions unless its applicable zero-retention controls are enabled. A provider’s general default does not prove a particular route’s configuration. Ask support about the provider and settings applicable to your release and request. Earlier zero-retention commitments remain applicable to requests made under them; this clarification does not authorize retaining those requests contrary to that commitment. Raw audio is handled locally or by Apple speech/transcription services where those services are used, not stored by our backend as a readable account record.

We do retain limited account, subscription, security, diagnostic, performance, and usage/accounting metadata, such as token counts, workflow labels, timestamps, processing duration, and derived cost. This metadata is used to enforce access limits, prevent abuse, operate the service, understand reliability and cost, and keep the service financially viable; it is not a record of what you said.

8.6 Direct contact data

Direct emails or website messages may remain until we no longer need them for follow-up, service-related replies, or legal/business recordkeeping, unless you ask us to delete them sooner and we are not legally required to keep them.

9. How Do We Keep Your Information Safe?

We use commercially reasonable technical and organizational measures, including where applicable:

No system can guarantee absolute security. Local-first storage may reduce some backend exposure compared with persistent server storage, but it does not eliminate risk if your unlocked device, exported files, backups, email account, or third-party provider accounts are compromised.

10. What Are Your Privacy Rights?

Depending on where you live, you may have rights to:

Because journal content is stored locally on your device after processing, you already control much of that data directly. For backend-linked account data or direct-contact data, contact us using the details below.

We do not currently use solely automated decision-making to make legal or similarly significant decisions about you in the sense used by privacy law. However, automatic entitlement, rate-limit, integrity, and safety checks may still enable, disable, delay, or block some app features.

If you are in the EEA, UK, Switzerland, Canada, or another jurisdiction that grants it, you may also have the right to complain to your local privacy or data protection regulator.

If processing depends on your consent, you can withdraw that consent at any time. This does not affect the lawfulness of processing before withdrawal or processing based on another lawful basis.

11. Advertising, Tracking, And Campaign Measurement

Voice of Self may be promoted through paid or organic campaigns on third-party platforms such as Meta, Instagram, TikTok, Apple Search Ads, Google, or similar services. We may choose campaign settings inside those platforms, such as region, language, placement, keyword, interest, app-category, campaign objective, budget, schedule, creative, and similar platform-provided targeting or optimization settings, where lawful.

For clarity, advertising Voice of Self on external platforms is different from showing ads inside the app. Voice of Self is not an ad-supported app: we do not display third-party ads, sponsored placements, ad-network content, or paid promotions inside the Voice of Self app.

If you see, click, view, save, share, comment on, install from, or otherwise interact with a Voice of Self ad on one of those platforms, the platform may process information about that interaction under its own terms and privacy policy. We may receive campaign reports from those platforms, such as impressions, clicks, installs, video views, engagement, spend, and other aggregated or campaign-level performance information.

We may use non-sensitive campaign links or parameters, such as UTM parameters, referral source, landing-page path, and aggregated platform reports, to understand whether a campaign is working. Website hosting, browser, and security systems may also receive ordinary request metadata such as IP address, user agent, request time, and the page requested.

Today, the Voice of Self app and website do not include Meta Pixel, TikTok Pixel, Google advertising tags, third-party ad SDKs, server-side conversion APIs, or similar ad-tracking code. We do not currently upload customer lists, account email addresses, support email addresses, device advertising identifiers, app activity, website activity, journal text, raw audio, transcripts, AI outputs, or similar user data to ad platforms for custom audiences, lookalike audiences, retargeting, attribution, or conversion measurement.

If we later add a website pixel, app event SDK, server-side conversion API, custom audience upload, retargeting audience, lookalike audience, or similar advertising measurement tool, we will treat that as a material advertising-data practice. Before activating that tool, we will update this policy and App Store privacy disclosures, use legally required cookie or tracking consent controls, and comply with Apple App Tracking Transparency where the iOS app tracks users across other companies’ apps or websites.

We will not intentionally send journal text, raw audio, transcripts, AI outputs, live-question content, direct-message text, children’s data, or other sensitive app content to Meta, TikTok, Google, Apple Search Ads, or similar ad platforms for targeting or campaign measurement. We also will not upload direct-contact lists or account email addresses for custom-audience advertising unless we have a lawful basis, required notice, and any required consent.

12. Controls For Do Not Track Features

Some browsers offer a Do Not Track (“DNT”) setting. Because there is no consistent industry standard for recognizing and enforcing DNT signals, we do not currently respond to them.

13. Do United States Residents Have Specific Privacy Rights?

If you live in a US state with a privacy law that applies to us, you may have rights to know, access, correct, delete, or obtain a copy of certain personal information, and to opt out of some kinds of processing where applicable.

13.1 Categories of personal information we may process

Depending on how you use the service, we may process categories such as:

These content categories are included here for privacy-law transparency. They do not mean that we retain a backend archive of what you said; normal managed AI request content is handled as described in section 8.5, separately from operational metadata and provider logging/retention.

Audio recordings may contain your voice, but we do not use recordings, transcripts, photos, images, or other content to identify people, create biometric identity templates, create voiceprints, perform speaker recognition, scan face geometry, or verify identity.

13.2 Sale, sharing, and sensitive information

We do not sell personal data. We do not currently share personal data from our app or website for targeted or cross-context behavioral advertising. We use sensitive personal information only as needed to provide the service, secure it, comply with law, or for other purposes permitted by applicable law.

If we add a practice that qualifies as sale, sharing, targeted advertising, or cross-context behavioral advertising under an applicable law, we will provide the legally required notice and opt-out or consent mechanism.

13.3 Exercising your rights

To exercise applicable rights, contact us using the details in Section 15.

13.4 Verification, authorized agents, and appeals

We may ask for information needed to verify your identity before responding to an access, correction, deletion, or similar privacy-rights request. Where applicable law allows, you may use an authorized agent to submit a request on your behalf, and we may ask for proof of that authorization.

If we deny a request and applicable law gives you an appeal right, you may appeal by emailing voiceofselfapp@gmail.com or by using the other contact details in Section 15.

13.5 Consumer health data note

Some U.S. laws use terms such as “consumer health data” for information that can identify or imply a person’s physical or mental health status, even outside traditional healthcare. Voice of Self is not healthcare, but entries, transcripts, reflections, or direct messages you choose to provide may include or imply health, mental-health, mood, grief, relationship, identity, or similar sensitive information.

We do not sell that information, use it for targeted advertising, share it for cross-context behavioral advertising, or use it to train our own models. We use it only as described in this policy: to provide features you request, operate and secure the service, respond to support you initiate, comply with law, and protect users or the service. If an applicable law requires a separate consumer health data notice, consent flow, or rights process for a market where Voice of Self is offered, we will provide the required notice or control before or with that processing.

We do not currently disclose personal information to third parties for their own direct-marketing purposes.

14. Do We Make Updates To This Policy?

Yes. We may update this Privacy Policy from time to time. The current version and effective date are published in the hosted legal manifest and surfaced in the app when a new acceptance is required.

15. How Can You Contact Us About This Policy?

If you have questions or comments about this Privacy Policy, you may contact us at:

16. How Can You Review, Update, Or Delete Data?

16.1 Local app data

You can delete local journals, transcripts, recordings, exports, and related content directly on your device. Standard .vos exports can include transcripts, episodes, AI notes, relics, and optionally settings. Removing the app may remove some local app data, but it may not remove copies in export files, Files/iCloud locations, email drafts, sent-mail folders, or device backups.

16.2 Backend-linked account data

Newer installation-identity releases: use the Support ID shown in Settings, where available, when emailing voiceofselfapp@gmail.com to request access or deletion of records tied to that identity. Save it before uninstalling. If you also used an older signed-in release, explain that so the older identity and backend can be reviewed separately. Do not send journal content or identity documents merely to open a request. We use proportionate verification and tell you if the information supplied cannot locate or establish control of records. These releases do not offer the legacy 30-day sign-in recovery procedure.

Where the app offers “Request access, correction or deletion,” it opens a draft that you review and send; it does not send a request automatically. It uses an existing or saved Support ID rather than creating an identity. If the ID is unavailable, contact support and explain the older account/release. Knowing an ID or restoring a purchase alone does not prove control of the older account. We request proportionate verification, not journal contents, passwords or identity documents by default.

Legacy releases with the in-app signed-in deletion flow: when completed, it immediately ends your existing sign-in sessions and removes these backend-linked records we control:

Two records are kept for a 30-day grace period and then removed automatically: your Firebase Authentication account record and your usage-accounting totals for the current billing period. Section 8.3.1 explains what that means, why it exists, and what those records are used for in the meantime — which is nothing else.

If you change your mind during those 30 days, signing back in with the same Apple or Google identity returns you to the same account. Note that the records deleted immediately are not restored by signing back in: your subscription status is re-established by restoring your purchase, and extra usage grants and stored files are gone.

If a backend, storage, or provider deletion step cannot be completed after retrying, the app leaves your sign-in account in place and fully usable, tells you which cleanup steps still need a retry, and lets you try the delete-account flow again later instead of silently finalizing account removal while cleanup is still incomplete. In that case no grace period starts, because the deletion did not take effect.

Deleting your account does not remove journals, transcripts, recordings, or other content that already exists only on your device. Deleting your account also does not automatically cancel an App Store subscription, and it does not necessarily remove transaction records, provider-side logs, platform logs, or legally required records kept by Apple, RevenueCat, Google, email providers, or other third parties under their own obligations.

16.3 Direct contact data

To request review, correction, or deletion of backend-linked data or direct-contact data, contact us using the details in Section 15. If you contacted us by email, you may also need to delete copies from your own mailbox, sent-mail folder, or other email systems you used.