Privacy Policy
Effective date: October 21, 2026
AccentPoof (“the app”) helps adults reduce their accent in English, one sound at a time. It is operated by the AccentPoof team (“we”, “us”). We are the controller of the personal data described here. Privacy contact: support@accentpoof.com.
AccentPoof is for people aged 18 and over.
Summary
- To score your pronunciation, the app records your voice when you tap the microphone. The audio is sent to our servers and to Microsoft’s speech service, which scores each sound. Most recordings are deleted automatically after 90 days.
- Your voice twin is optional. We make it only after you agree in the app and confirm you are 18 or older. It is used only to play your own sentences back to you with a lighter accent. You can delete it at any time.
- We do not use your recordings to train AI models, we do not sell your personal information, and we do not share it for advertising. The app has no ad SDKs and does not read your advertising identifier.
- We do not record your screen. Session replay is off.
- Payments are handled by Apple or Google. We never see your card details.
- You can delete your account and all your data at any time: in the app (Me > Account > Delete account) or by following Delete your account.
One login across our apps
We offer more than one app, and our apps share one sign-in system. If you use the same Apple, Google or email sign-in in another one of our apps, it is the same login. Each app keeps its own data separate and only that app can read it. When you delete your AccentPoof account, we delete your AccentPoof data. Your underlying login is deleted too, unless you also use another one of our apps. In that case the login stays so that app keeps working.
When you start
When you open the app for the first time, it creates an anonymous account on our servers so it can score your 60 second check. No name or email is needed for that. If you save an account later (Apple, Google or an email code), your progress moves to it.
Data we collect
| Data | What it includes | Why | How long we keep it |
|---|---|---|---|
| Voice recordings | audio of the lines and words you say when you tap the microphone, the line you were asked to read, and the length of the take | score your sounds, build and adjust your plan, let you replay your takes | audio: deleted 90 days after recording, except the recordings described in “Recordings we keep longer” |
| Scores and transcripts | how each sound, word and line scored, what the speech service heard, and your progress per sound | show your sound map, plan your lessons, show your progress | until you delete your account |
| Voice twin (optional, with your consent) | short audio clips of your sentences in your own voice with a lighter accent, and the record of your consent | let you hear the target in your own voice | see “Voice twin” below |
| Account data | user ID, email, your name if Apple or Google shares it, a display name | create and secure your account, sync your devices | until you delete your account |
| Profile and plan | first language, interface language, goals, work field and role you choose, interests, words you add, target accent, daily minutes, reminder time, time zone, country | personalize your plan, tips and practice words | until you delete your account |
| Learning activity | lessons, practice days, streaks, points, words to review, monthly accent checks, invite links you share | run the daily plan and show progress | until you delete your account |
| Reports and support messages | the reason and note of a content report, your support email | fix content and answer you | until you delete your account |
| Subscription data | product, period, store, renewal state | unlock Pro and restore purchases | until you delete your account; transaction records as the law requires |
| Usage data | screens viewed, buttons used, onboarding answers, paywall and purchase events, an analytics ID linked to your user ID. No audio and no text you type | understand which features work and fix problems | see Retention |
| Diagnostics | crash logs, device model, system and app version; on our servers, your user ID with an error | fix bugs | see Retention |
| Approximate location | country from your device region setting, and country or region derived from your IP address by our servers and providers | language packs, prices, regional rules, security | as the data above |
| Install source | the Apple Search Ads attribution token and campaign tags from the link that led to the install, if any | learn which channels bring people to the app | see Retention |
We do not collect precise location, contacts, photos or your advertising identifier. Reminders are notifications scheduled on your phone; we do not run a push server.
Voice recordings
The app uses the microphone only while you are recording, after you allow it in the system prompt. Your recordings are stored in private storage that only our server can open, through short-lived links.
Who processes them. To score a recording, our server sends the audio and the line you were reading to Microsoft Azure AI Speech (United States, East US region), which returns scores for each sound and word. Microsoft processes the audio on our behalf under its terms for Azure services. We do not send your name or email with a recording.
Recordings we keep longer. Your day-1 challenge recordings and your best take of each line are kept until you delete them or your account, so you can compare Day 1 with today. If you choose to make a voice twin, the day-1 recordings are also its reference. Day-1 recordings are deleted at the latest 3 years after your last use of the app.
We do not use your recordings to train AI models, and we do not sell or share them.
Delete them. Me > Recordings > Delete all my recordings deletes your recordings from your phone and from our servers. Deleting your account deletes them too.
Voice twin
The voice twin lets you hear your own sentences in your own voice with a lighter accent. It is optional, and you can use AccentPoof without it.
- Consent first. The app explains the voice twin and asks you to check two boxes: that you agree to create it from your recordings, and that you are 18 or older. We store that consent with the date, the text version, the language and the app version. Without it, nothing is created.
- How it is made. Microsoft’s text to speech reads your sentences in the target accent. Then our own voice service, which runs an open-source voice conversion model on servers we rent from Modal (United States), changes the sound of that reading to match your voice, using your day-1 recordings as the reference. The service keeps nothing between jobs: the voice features it computes and every temporary file are deleted when the job ends. The clips carry an inaudible watermark that marks them as AI generated.
- What we store. The finished clips, in private storage, and your consent record.
- What we never do. We never use your voice twin for anything except playing your practice back to you. We never sell it, share it with other users, use it in marketing, or use it to train other voices.
- Retention schedule. We destroy your voice twin clips and the day-1 recordings used as its reference at the earliest of: when you delete your voice twin, when you delete your account, or 3 years after your last use of the app. Voice twin deletion runs right away; if a step fails, our daily cleanup finishes it within 24 hours.
- Delete it. Me > Voice twin > Delete my voice twin, or delete your account.
Some laws treat voice data used this way as biometric or sensitive information, for example in Illinois, Texas and Washington, and as sensitive data under Brazil’s LGPD. We ask for your explicit consent, use the data only for the purpose above and follow the retention schedule above.
How we use data and legal bases
- Provide the service (scoring, plan, lessons, account, sync, subscriptions): performance of a contract (LGPD art. 7, V; GDPR art. 6(1)(b)).
- Voice twin: your consent (LGPD art. 11, I; GDPR art. 9(2)(a) where it applies; written release under Illinois BIPA).
- Security, abuse prevention, crash diagnostics: legitimate interests (LGPD art. 7, IX; GDPR art. 6(1)(f)).
- Product analytics (no audio, no typed text): legitimate interests. You may object at support@accentpoof.com.
- Legal obligations (tax, accounting, lawful requests): LGPD art. 7, II; GDPR art. 6(1)(c).
Service providers
We use these providers under contracts that limit what they can do with your data.
| Provider | Purpose | Location |
|---|---|---|
| Supabase | database, file storage, sign-in | United States (AWS us-east-1) |
| Vercel | runs our app server and hosts this website | United States (East) |
| Microsoft Azure AI Speech | scores your recordings; text to speech for Mo’s voice and the voice twin source | United States (East US) |
| Modal | runs our voice twin service | United States |
| RevenueCat | subscription management, install attribution | United States |
| PostHog | product analytics (session replay is off) | United States (PostHog US Cloud) |
| Sentry | crash and error reporting | United States |
| Resend | sends the email with your sign-in code | United States |
| Apple | Sign in with Apple, App Store distribution and payments | worldwide |
| Sign in with Google, Google Play distribution and payments | worldwide |
International transfers
Our servers and providers are in the United States. If you live outside the US, for example in Brazil, Mexico or Canada, your data is transferred to the US when you use the app. We rely on the safeguards available under your law, including the mechanisms of LGPD art. 33 and standard contractual clauses where they apply. In the United States, your data is subject to US law and may be accessed by US courts and authorities under that law.
Retention
- Recording audio: deleted 90 days after recording, except your day-1 recordings and best takes, which stay until you delete them or your account (day-1 recordings at the latest 3 years after your last use of the app).
- Voice twin: see the retention schedule in “Voice twin”.
- Your other AccentPoof data (profile, scores, plan, progress, reports, subscription record): kept until you delete your account.
- Items that can outlive deletion: encrypted backups, analytics and diagnostics records at our providers, and transaction records the law requires us to keep. We keep these only as long as needed to provide the service or meet legal obligations, then delete or anonymize them.
- Apple and Google keep their own purchase records under their own policies.
Your rights
- Everyone: access, correct, delete and receive a copy of your data, withdraw consent, and contact support@accentpoof.com. We answer within 30 days.
- Brazil (LGPD): confirmation of processing, access, correction, anonymization, blocking or deletion, portability, information about sharing, information about the possibility of refusing consent, and revocation of consent (art. 18). You may petition the ANPD. Our data protection officer (encarregado): AccentPoof privacy team, support@accentpoof.com.
- Mexico (LFPDPPP): access, rectification, cancellation and opposition (ARCO rights), and revocation of consent, by writing to support@accentpoof.com.
- Canada (PIPEDA and provincial laws, including Quebec’s Law 25): access and correct your personal information, withdraw consent, ask how your information is handled outside Canada, and complain to the Office of the Privacy Commissioner of Canada or your provincial regulator. The person responsible for the protection of personal information is the AccentPoof privacy team, support@accentpoof.com.
- California (CCPA/CPRA): right to know, delete, correct, and to limit the use of sensitive personal information. We do not sell or share personal information. We use sensitive personal information (voice recordings, the voice twin, account sign-in) only to provide the service you ask for. Categories collected in the last 12 months: identifiers, commercial information (purchases), internet or electronic activity, audio information, inferences about your pronunciation, and approximate location. We do not discriminate against you for using your rights.
- Other US states: several states have privacy and biometric laws. In every state we create a voice twin only with your consent, never sell voice data, and delete it on request and on the schedule above.
- EEA and UK: if the GDPR or UK GDPR applies to you, you also have the right to restrict or object to processing, to data portability, and to complain to your data protection authority.
Children
AccentPoof is for people 18 and over. We do not knowingly collect data from anyone under 18. If you believe a minor gave us data, write to support@accentpoof.com and we will delete it.
This website
This website does not use analytics or advertising cookies. If you pick a language, we store one cookie named lang to remember it. The share buttons on our guides are plain links: nothing is sent to a social network until you click one. Our hosting provider keeps standard server logs, which include IP addresses.
Security
Data is encrypted in transit (TLS) and at rest by our hosting providers. Recordings and voice twin clips sit in private storage that only our server can open. Each account can read only its own data. No system is perfectly secure. If a breach affects you, we will notify you and the authorities as the law requires.
Changes
We will post updates on this page and change the date above. If a change is material, we will tell you in the app, and we will ask again before using your voice data for anything new.
Contact
The AccentPoof team. Privacy and support: support@accentpoof.com.