Accentually Privacy Policy

Preserving regional accents — with your trust

Effective date: 1 April 2026

1. Who We Are

Accentually is a mobile application designed to capture, preserve, and celebrate regional accents, dialects, sayings, and slang from across the United Kingdom and beyond. It is developed by Hugh Mackenzie ("we", "us", "our"), based in the United Kingdom.

We are the data controller for any personal data processed through Accentually. If you have any questions about this policy or your data, you can contact us at:

2. Our Mission & Why We Collect Data

Every generation loses a little of how their grandparents spoke. Regional accents, sayings, and slang are disappearing as communities change and media homogenises language. Accentually exists to record and preserve this linguistic heritage before it fades away.

Your voice recordings and the phrases you share help us build one of the richest collections of regional speech data anywhere. This data serves three purposes:

1. Preservation: Creating a living record of how real people actually speak across the UK
2. AI training: Teaching AI models to accurately recognise, understand, and reproduce regional accents — so future voice assistants can sound like a real person from your area, not a generic robot
3. Research: Contributing to linguistic and cultural research about regional speech patterns

3. What We Collect

3.1 Account data

Data	Purpose	Storage
Email address (via Apple Sign In or Google Sign In)	Account authentication and communication	Cloudflare D1 (EU region)
Display name	Identifying you within the app (leaderboards, voting)	Cloudflare D1
Claimed region/accent	Associating recordings with the correct accent region	Cloudflare D1
Consent preferences	Recording which consents you have granted or withdrawn	Cloudflare D1

3.2 Voice recordings

When you record yourself speaking phrases or have a voice conversation with our AI, we capture audio recordings of your voice. Voice recordings are:

• Stored as M4A/AAC format (64kbps mono) on Cloudflare R2 (EU region)
• Converted to WAV 16kHz for AI processing on Modal.com
• Associated with your account and your claimed accent region
• Used for accent scoring, AI training, and linguistic research (depending on your consent level)

3.3 AI conversation data

When you use the voice or text conversation feature with our AI accent researcher:

• Your messages (text transcripts from speech-to-text, or typed messages) are sent to Anthropic's Claude API for generating responses
• The phrases, sayings, and slang you share in conversation may be added to our dialect database
• Conversation history is maintained for the duration of your session to provide context
• We do not store complete conversation transcripts long-term; we extract and retain only the linguistic data (phrases, dialect features, regional expressions)

3.4 Voting and interaction data

Data	Purpose	Storage
Hot or Not votes on other users' recordings	Community accent rating feature	Cloudflare D1
Leaderboard scores and points	Gamification and engagement	Cloudflare D1
Referral codes	Tracking referral rewards	Cloudflare D1

3.5 Data we do NOT collect

Precise location / GPS data	Contacts or address book
Advertising identifiers (IDFA/GAID)	Browsing history
Financial or payment data (handled by Stripe)	Photos or camera roll access
Phone number	Social media profiles

4. Biometric Data Notice

4.1 Two-tier consent model

Accentually uses a clear two-tier consent system. You choose exactly how your voice data is used:

4.2 Consent for AI conversations

When you use the voice conversation feature, your speech is transcribed using on-device speech recognition (Apple SFSpeechRecognizer on iOS, Google SpeechRecognizer on Android). The text transcript (not your raw audio) is sent to Anthropic's Claude API. The AI's spoken responses are generated via our text-to-speech service.

Phrases, sayings, and dialect features you share during conversation may be extracted and added to our dialect database to enrich the linguistic data available for your region. This data is associated with a region, not with your personal identity.

4.3 Proxy recordings (recording someone else)

You may record someone else's voice with their verbal permission. When you create a proxy recording:

• The person being recorded does not need to create an account
• We do not collect any personal information about the recorded person (no name, email, or account data)
• A unique deletion token (a code starting with ACC-DEL) is generated and must be given to the recorded person
• The recorded person can use this token to request deletion of their recording at any time, without needing an account
• You are responsible for obtaining the recorded person's verbal consent before recording them

4.4 Retention and destruction of biometric data

• Voice recordings used only for accent scoring (Tier 1) are retained for the lifetime of your account
• Voice recordings used for AI training (Tier 2) are retained for up to 5 years from the date of collection
• When you delete your account, all associated voice recordings are permanently deleted within 30 days
• Proxy recordings are deleted upon receipt of a valid deletion token request
• Truly anonymised acoustic features extracted from recordings (which cannot identify any individual) may be retained beyond deletion

4.5 Your right to refuse

You can use Accentually's text conversation, voting, and leaderboard features without ever recording your voice. Voice recording is not required for basic app functionality. You can withdraw Tier 2 (AI training) consent at any time without affecting your ability to use the app.

5. How We Use Your Data

5.1 Accent scoring and community features

We process your voice recordings through our AI accent classifier (a wav2vec2-based model running on Modal.com) to score how closely your speech matches your claimed regional accent. Your recordings may also be made available for community voting (Hot or Not) with your consent.

5.2 AI conversation (Anthropic Claude)

When you use the AI conversation feature, text transcripts of your messages and conversation context are sent to Anthropic's Claude API (specifically Claude Haiku) to generate responses. Your voice audio is not sent to Anthropic — only the text transcript from on-device speech recognition. Anthropic processes this data under their privacy policy and does not use API inputs to train their models.

5.3 AI model training

With your explicit Tier 2 consent, your voice recordings are used to train machine learning models that can:

• Recognise and classify regional accents accurately
• Generate authentic text-to-speech in regional accents (future feature)
• Detect accent features such as gender, age range, and dialect intensity

5.4 Dialect database enrichment

Phrases, sayings, slang, and dialect features shared during conversations or recordings are added to our dialect phrase database. This data is associated with a region, not with your personal identity. For example, we may record that "howay" is used as a greeting in Newcastle, but we do not record that you specifically said it.

5.5 Data licensing (future)

We may license anonymised voice and accent datasets to AI voice companies (such as ElevenLabs, Google, Amazon), academic institutions, and research organisations. This data is used to improve how AI handles regional accents. See Section 6 for how we anonymise data. We will never sell your personal information or identity.

6. Anonymised & Aggregated Data

We create anonymised datasets from voice recordings by:

• Extracting acoustic features (pitch, rhythm, vowel formants) as numerical data, discarding the original audio
• Removing all account information, email addresses, and personal identifiers
• Aggregating data across many speakers so individual voices cannot be distinguished
• Retaining only the regional accent label (e.g. "Newcastle", "Glasgow") alongside the acoustic data

Once data has been truly anonymised such that re-identification is not reasonably possible, it is no longer considered personal data under applicable data protection laws. We may use anonymised data for any lawful purpose, including AI training, research, and commercial licensing.

We commit to not attempting to re-identify any individual from anonymised data, and we contractually require the same of any third party to whom we provide datasets.

7. Who We Share Data With

Recipient	Data shared	Purpose
Anthropic	Text transcripts of conversations (no raw audio, no email)	AI conversation responses (Claude Haiku)
Modal.com	Voice recordings (audio files)	AI accent scoring and model training (GPU inference)
Cloudflare	Account data, voice recordings, app data	Infrastructure: database (D1), file storage (R2), caching (KV), API hosting (Workers)
Apple / Google	Authentication tokens only (via Sign In with Apple / Google Sign In)	Account creation and login
Stripe (future)	Payment information (handled entirely by Stripe, never touches our servers)	Subscription payments for premium features
Research partners & AI companies (future)	Anonymised datasets only	AI voice model training, linguistic research

We do not sell personal data. We do not share data with advertisers. We do not use data for profiling, targeted advertising, or any purpose unrelated to accent preservation and AI training.

8. Data Security

We take data security seriously. Your data is protected by:

• Transport encryption: All data is transmitted over HTTPS/TLS
• Authentication: API requests are authenticated using session tokens with expiry
• Storage: Voice recordings are stored on Cloudflare R2 with access controls. Account data is stored in Cloudflare D1 (EU region, Western Europe)
• AI processing: Voice files sent to Modal.com for inference are processed in-memory and not persisted beyond the inference request (for scoring). Training data is stored on encrypted Modal volumes
• Access control: Only authorised personnel (the developer) have access to production data
• Password hashing: If you use email/password login, passwords are hashed using bcrypt and never stored in plaintext

9. Data Retention

Data type	Retention period
Account data (email, display name, region)	Lifetime of your account, deleted within 30 days of account deletion
Voice recordings (Tier 1: scoring only)	Lifetime of your account
Voice recordings (Tier 2: AI training)	Up to 5 years from date of collection
Proxy recordings	Until deletion token is redeemed or account holder deletes them
AI conversation transcripts	Session duration only — not stored long-term
Dialect phrases extracted from conversations	Retained indefinitely (associated with region, not personal identity)
Anonymised acoustic features	Retained indefinitely (no longer personal data)
Consent records	7 years (legal compliance)
Deletion request records	7 years (legal compliance)

10. Lawful Basis for Processing (UK Users)

Under the UK General Data Protection Regulation (UK GDPR), we process your data on the following lawful bases:

Processing activity	Lawful basis
Account creation and management	Contract (Art. 6(1)(b)) — necessary to provide the service
Voice recording for accent scoring	Explicit consent (Art. 9(2)(a)) — biometric data requires explicit consent
Voice recording for AI training	Explicit consent (Art. 9(2)(a)) — separate optional consent
AI conversation (sending text to Anthropic)	Consent (Art. 6(1)(a)) — you choose to use the conversation feature
Community features (voting, leaderboards)	Legitimate interests (Art. 6(1)(f)) — necessary for app functionality you expect
Dialect phrase extraction from conversations	Legitimate interests (Art. 6(1)(f)) — core purpose of the app, data is anonymised to region level
Data licensing (anonymised datasets)	Not applicable — truly anonymised data is not personal data under UK GDPR

You may withdraw your consent at any time through the app's account settings. Withdrawal of consent does not affect the lawfulness of processing carried out before withdrawal.

11. Your Rights (UK)

Under the UK GDPR, you have the following rights:

• Right of access — request a copy of the personal data we hold about you
• Right to rectification — request correction of inaccurate data
• Right to erasure — request deletion of your personal data (see Section 14)
• Right to restrict processing — request that we limit how we use your data
• Right to data portability — receive your data in a machine-readable format
• Right to object — object to processing based on legitimate interests
• Right to withdraw consent — withdraw consent at any time via the app or by contacting us
• Rights related to automated decision-making — the accent score is generated by AI but has no legal or similarly significant effect on you

To exercise any of these rights, contact us at privacy@accentually.app. We will respond within one month.

You also have the right to lodge a complaint with the Information Commissioner's Office (ICO):

12. Your Rights (US Users)

California (CCPA/CPRA)

If you are a California resident, you have the right to:

• Know what personal information we collect and how it is used
• Delete your personal information
• Correct inaccurate personal information
• Opt out of the sale or sharing of personal information — we do not sell personal information
• Limit use of sensitive personal information — voice recordings are sensitive PI; you control collection via the consent tier system
• Non-discrimination — we will not discriminate against you for exercising your rights

Illinois (BIPA)

If you are an Illinois resident, please see Section 4 (Biometric Data Notice) for specific disclosures required under the Illinois Biometric Information Privacy Act, including our collection practices, purpose, retention schedule, and your right to refuse.

Other US States

Residents of Virginia, Colorado, Connecticut, Texas, Oregon, Washington, and other states with comprehensive privacy laws have similar rights to access, delete, correct, and opt out. Contact us at privacy@accentually.app to exercise your rights.

13. Children's Privacy

Accentually is intended for users aged 13 and over. We do not knowingly collect personal information from children under 13.

If you are between 13 and 16 years old and resident in the UK or EU, you need your parent or guardian's consent to create an account and record your voice.

If we become aware that we have collected data from a child under 13, we will delete it promptly. If you believe a child under 13 has used Accentually to record their voice, please contact us at privacy@accentually.app.

14. How to Delete Your Data

In the app

• Open Account Settings
• Tap "Delete My Account" to permanently delete your account and all associated recordings
• You can also delete individual recordings from your profile
• You can withdraw Tier 2 (AI training) consent at any time without deleting your account

Proxy recording deletion (deletion tokens)

• If someone recorded your voice using Accentually, they should have given you a deletion token (a code starting with ACC-DEL)
• Email the token to privacy@accentually.app and we will delete the recording within 30 days
• You do not need an account to request proxy recording deletion

By email

Contact privacy@accentually.app with your account email address. We will process deletion requests within 30 days.

What happens when you delete:

• Your account data (email, display name, preferences) is permanently erased
• All your voice recordings are permanently deleted from Cloudflare R2
• Your votes, scores, and leaderboard entries are removed
• Anonymised acoustic features and dialect phrases (which cannot identify you) may be retained
• If your voice data was included in a licensed dataset, we will notify the licensee of the deletion requirement

15. International Data Transfers

Your data is primarily stored and processed in the following locations:

Service	Location	Data
Cloudflare D1, R2, KV	EU (Western Europe)	Account data, voice recordings, app data
Modal.com	United States	Voice recordings (for AI processing)
Anthropic (Claude API)	United States	Text conversation transcripts

Where data is transferred outside the UK, we rely on appropriate safeguards including Standard Contractual Clauses (SCCs) and the service providers' data protection commitments. Cloudflare, Anthropic, and Modal.com each maintain comprehensive data protection practices.

16. Changes to This Policy

We may update this privacy policy from time to time. If we make material changes to how we collect or use your data, we will:

• Update the "Effective date" at the top of this page
• Display a notice within the app on first launch after the change
• If the change relates to biometric data processing or consent tiers, request your consent again

We encourage you to review this policy periodically.

17. Contact Us

We aim to respond to all privacy enquiries within 30 days.