Privacy Policy
Last updated: 23 October 2026
This Privacy Policy describes how Parlilo ("Parlilo", "we", "us", or "our") collects, uses, shares, and protects information when you use the Parlilo mobile app or this website (together, the "Service").
We aim to collect as little personal data as possible and to be transparent about what we do collect. If you have questions, email support@parlilo.com.
Optional website analytics
When enabled for this website, Google Analytics runs only after you choose Accept analytics. It measures marketing page visits, pricing visits, sample lesson links, campaign language selections and app-store clicks so we can understand which parts of the website are useful. We send known page paths and labels such as language, button location and destination store. For recognised campaigns we also send fixed campaign, source, medium and creative labels; we do not send account-action codes, URL query strings, conversation text or account identifiers in these events. Google Analytics uses cookies and processes technical information about your browser and device. Advertising features are disabled.
Your choice is stored in this browser. Use Analytics preferences in the footer to change it. Rejecting analytics prevents collection; withdrawing permission stops future collection and removes this site's Google Analytics cookies. Analytics is not required to use the website. See Google's information about partner-site data for details about its processing.
Who we are
Parlilo is published by Hatch Software, the publisher of the dev.voca.app mobile application. This policy applies to the app available on the App Store and Google Play, and to the marketing site at this domain.
What we collect
We collect the following categories of information when you use the Service:
- Account information. When you sign in, Firebase Authentication stores your email address or sign-in provider identifier and issues a unique user ID. Depending on your sign-in method, we also receive your display name and profile-picture URL.
- Learning data. Your selected target language, native language, completed scenarios, vocabulary progress (spaced repetition state), session scores, and CEFR estimates.
- Voice recordings, transcripts and messages. We capture audio when you use a recording or dictation feature, and receive text when you type or submit a speech transcript. Our speech and AI providers process this content as described under Speech and AI processing below. We store conversation transcripts, results and Tutor recaps with your account. Some session audio is also stored for pronunciation assessment and review.
- Subscription and purchase data. RevenueCat receives your purchase events from Apple and Google so we can validate your entitlement to Parlilo Pro. We receive a customer ID, entitlement status, expiry date, and store of purchase.
- Device and diagnostic data. Crash reports, error logs, device metadata (OS, app version, locale), and session and activity identifiers help us debug problems. Crash reports can be linked to your account ID. Server logs can also contain reference phrases, recognised speech and provider error details when processing or diagnosing a speech request.
- Usage analytics. Events about which screens you visit and which scenarios you complete, so we can understand what's working. The app sends a pseudonymous account identifier, and Firebase Analytics uses app-instance identifiers and derives approximate location from masked IP addresses. These analytics can be associated with your account; they do not contain your conversation text. On Android, recognised campaign labels supplied by Google Play’s install referrer help us understand whether a campaign leads to learning and purchases. Raw referrers are not retained in our campaign measurement.
We do not access your contacts, device photo library, device location services, advertising IDs, or browsing history. A sign-in provider may supply a profile-picture URL as described above. Approximate location inferred by Firebase Analytics from masked IP addresses is described above.
How we use it
- To run the Service — deliver voice sessions, score your speaking, track your progress.
- To manage your subscription and prevent abuse of the free tier.
- To diagnose crashes and improve quality.
- To respond to you when you contact support.
We do not use your data to train third-party advertising models, and we do not sell it.
Speech and AI processing
The information shared depends on the feature you use:
- Spoken conversations and Tutor. Your device streams microphone audio directly to ElevenLabs for speech recognition. The resulting transcript, or a message you type instead, is sent through Parlilo's server to OpenAI with the conversation history and teaching context, such as your target language, native language, level, exercise, hints and learning objectives. OpenAI generates replies, coaching and feedback. Parlilo's server sends reply text to ElevenLabs to generate spoken audio; reply text may repeat information you supplied during the conversation.
- Speaking drills and conversation pronunciation. Recorded audio, the reference phrase or transcript and the assessment language are sent through Parlilo's server to Microsoft Azure Speech for pronunciation assessment. Conversation audio may be uploaded automatically after a spoken session finishes so we can produce feedback. Typing a conversation without recording audio does not create a pronunciation recording.
- Corrections and text assessment. OpenAI receives the original phrase or transcript, corrected text, explanation, exercise prompt and answer as needed, together with your languages and level, to create practice and assess your response. A transcript-based assessment sends text rather than the original recording to OpenAI.
- Dictating a correction-practice answer. This feature uses your device's speech-recognition service. Depending on your device and its configuration, audio may be processed by Apple on iOS or the configured speech-recognition provider on Android, such as Google. Processing is not guaranteed to stay on the device. The recognised answer is sent to OpenAI when you submit it for assessment.
- Listening to phrases and examples. Authored lesson text, example answers, corrections or other text you ask to hear may be sent to ElevenLabs to generate audio when a cached clip is unavailable. Merely listening to an authored lesson phrase does not send a microphone recording.
We do not add your account email address, password or payment details to these AI prompts. Anything personal you include in speech or text can nevertheless become part of a transcript, prompt or spoken reply. Providers also receive technical information needed to handle the connection; ElevenLabs receives your device's network connection directly for live transcription.
On iOS, we ask you to agree to this processing before entering the learning experience. You can read this policy first or sign out without agreeing. Acceptance is saved for your account on that device; reinstalling the app or a materially revised notice requires agreement again.
If you do not want to send new speech or text for this processing, do not start or submit the corresponding recording, dictation, conversation, correction or assessment feature. You can revoke microphone and speech-recognition access in your device settings. Revoking microphone access does not prevent typed messages from being processed, remove data already sent, or delete your account history. Contact us to request deletion or exercise your data-protection rights.
How we share it
We share data only with the service providers that make the Service work:
- Google Firebase and Google Cloud — authentication, account and learning records in Firestore, stored pronunciation audio and generated audio clips, API and website hosting, analytics, crash reports and server logs. See Firebase's privacy and security information.
- OpenAI — the conversation, coaching, correction and text-assessment processing described above. See OpenAI's API data controls.
- ElevenLabs — the speech-recognition and speech-generation processing described above. Its Data Processing Addendum describes processing of customer personal data; training controls and retention information describe available account and service options.
- Microsoft Azure Speech — the pronunciation assessment described above. See Microsoft's Speech data privacy information.
- RevenueCat — your app user ID, purchase and subscription information used to manage entitlement to Parlilo Pro. See RevenueCat's privacy policy.
- Apple and Google — payment processing through the App Store and Google Play and, where used, sign-in and device speech recognition. See Apple's privacy policy and Google's privacy policy.
We share data with service providers as needed to operate the Service, under the applicable service agreements and data-processing terms. Provider retention and model-improvement practices depend on the service, contract and account settings; using an API does not by itself mean that a provider retains no data or never uses data to improve its models.
Providers that process personal data on our behalf must provide the same or equal protection of that data as described in this policy and required by applicable data-protection law. Their applicable data-processing terms require safeguards for confidentiality and security, limit processing to the agreed purposes and legal obligations, and provide for assistance with data-protection requests. See the OpenAI Data Processing Addendum, ElevenLabs Data Processing Addendum, Firebase Data Processing and Security Terms and Microsoft Data Protection Addendum.
We may also disclose data if required by law, to protect our rights, or as part of a corporate transaction (acquisition, merger), in which case you will be notified.
Retention
- Voice recordings: conversation clips uploaded for pronunciation assessment can be stored in Google Cloud under your account until account deletion. The current audio bucket has no automatic age-based deletion rule. Deleted clips can remain recoverable in Google Cloud's soft-deletion storage for seven days. Speaking-drill recordings are sent for assessment without being saved by that drill endpoint to our audio bucket. Speech-provider retention is separate from Parlilo's storage.
- Generated phrase audio: audio generated for phrase playback is cached separately from your account. A clip can contain personal information if the text being spoken includes information you supplied. These cached clips are not automatically removed by account deletion; contact us to request removal of a clip containing your personal information.
- Transcripts and session scores: retained for as long as your account exists, so you can review history.
- Account data: retained while your account is active.
- Diagnostics and security logs: application server logs in our default Google Cloud logging bucket are retained for 30 days. Firebase states that Crashlytics retains crash reports and associated identifiers for 90 days before beginning deletion from live and backup systems. Required Google Cloud audit logs are retained for 400 days for security and accountability.
- Provider records: OpenAI documents that API content is not used for model training by default and that abuse-monitoring records may be retained for up to 30 days, with exceptions described in its API data controls. Microsoft states that it does not retain input to its real-time pronunciation-assessment service. ElevenLabs retention and model-improvement use depend on its applicable terms and account settings; zero retention is an optional configuration for eligible services and customers.
Deleting your account
You can delete your account from inside the app under Settings → Account → Delete account, or by emailing support@parlilo.com. Deletion removes your account record, learning history, and subscription metadata from our systems within 30 days, except where we are required to retain certain records (for example, financial records).
We retain a minimal record of the deleted user ID and deletion time to prevent outstanding authentication tokens or later subscription events from recreating your deleted account. Deleting a Parlilo account does not cancel an App Store or Google Play subscription, and does not itself erase records independently retained by Apple, Google, RevenueCat or speech and AI providers under their applicable terms. Contact us for help with requests concerning data shared through Parlilo.
Children
Parlilo is not directed at children under 13. If you are under 13, please do not use the Service. If you believe a child has provided us with personal data, email us and we will delete it.
Your rights
Depending on where you live (UK, EU, California, and others), you may have rights to access, correct, export, or delete your personal data, and to object to certain processing. To exercise these rights, email support@parlilo.com and we will respond within the statutory window.
Security
We use HTTPS for all network traffic, encryption at rest for our databases, and least-privilege access for our team. No service is perfectly secure, but we work hard to keep yours safe.
International transfers
Parlilo processes data in the United Kingdom, the European Union, and the United States. Where personal data is transferred outside your country, we rely on standard contractual clauses or equivalent mechanisms required by law.
Changes to this policy
If we make material changes, we will notify you in the app or by email before the changes take effect. Minor changes (clarifications, formatting) will be reflected by updating the "last updated" date above.
Contact
Questions or requests? Email support@parlilo.com.