Skip to content

Privacy policy

Version 1.0 · effective November 7, 2026

Document language: English · Русский

Contents

1. About this policy

  1. 1.1

    This policy explains how the operator of Frazo collects, uses and protects personal data when you use the website frazo.app, the author dashboard and the Telegram bot @frazo_bot.

  2. 1.2

    Frazo helps experts write from their own words: it interviews you by voice, transcribes your answers and turns them into drafts of posts, carousels, Reels scripts and articles.

  3. 1.3

    The operator is the controller of your personal data and is named in the section Operator and contacts. Questions about your data: hello@frazo.app.

  4. 1.4

    This policy applies to users of the English version of the site. Users of the Russian version are covered by the Russian documents at frazo.app/ru/docs.

Back to contents

2. What data we process

2.1 Account data

Your email address. If you sign in with Telegram: your numeric Telegram ID and the ID of your chat with the bot. The name you give in your profile.

2.2 Profile and settings

Your answers to the onboarding questions and quiz, your niche and topics, a description of how you write and speak, sample texts if you send them. Interface and post language, time zone, reminder and publishing schedule, retention period, connected Telegram channels.

2.3 Recordings and texts

Voice recordings made on the site and voice messages sent to the bot. Typed answers. Transcripts with timestamps. Interview topics and angles, drafts and their versions, revision notes, accepted and published texts, memory facts with quotes, and your speech profile: pace, frequent words and habitual phrases.

2.4 Technical data

IP address, browser and device details of your sign-in session, cookies, the country derived from your IP address, service events without any text (for example, that an answer was recorded or a draft accepted), server logs. In the consent log: the date, document version, IP address and a pseudonym of your email, which is a one-way fingerprint that cannot be turned back into the address.

2.5 Visitors of the site

On the English version of the site, Google Analytics runs only if you accept it in the cookie banner. See the Cookie policy.

2.6 What we do not process

We do not receive or store payment card data: there are no payments in the service yet. We do not identify you by your voice and do not use it as a password. We do not aim to process health data or other special categories of data; if you mention such data in a recording yourself, it is processed as part of that recording and deleted with it.

Back to contents

3. Why we use your data and on what legal basis

  1. 3.1

    We use your data only for the purposes below. Where the GDPR or UK GDPR applies, the legal bases are those of Article 6(1).

Purposes and legal bases

Purposes and legal bases
PurposeDataLegal basis
Creating and running your account, signing you inEmail, Telegram ID, session, IP addressPerformance of our contract with you (b)
Interviews, transcription, drafts, memory, posting to your channelProfile, recordings, transcripts, texts, memoryPerformance of our contract with you (b) and your consent (a)
Daily topics, reminders and service messages in Telegram and by emailEmail, Telegram ID, schedulePerformance of our contract (b); you can switch them off in Settings
Data export, deletion and answering your requestsAll account dataLegal obligation (c)
Keeping proof of your consent and its withdrawalConsent logLegal obligation (c)
Protecting the service from abuse: rate limits, protection against code guessingIP address, email pseudonyms, eventsOur legitimate interest in a secure service (f)
Improving the service from events without textService eventsOur legitimate interest in understanding how the service is used (f)
Visit statistics for the English siteGoogle Analytics dataYour consent (a), given in the cookie banner
Back to contents

4. How drafts are made

  1. 4.1

    Drafts are generated by a language model from your recordings. The service checks each sentence against your recording, shows its source and flags sentences that are not in the recording.

  2. 4.2

    We make no automated decisions that have legal or similarly significant effects on you. A text is posted to your channel only after you have accepted it and chosen the time.

  3. 4.3

    Frazo has no models of its own and does not use your recordings or texts to train models. Our requirements for model providers are described below.

Back to contents

5. Where your data is stored

  1. 5.1

    All service data is stored on a server the operator rents from Deluxhost in the Netherlands: the database, voice recording files, data archives and database backups.

  2. 5.2

    Voice messages you send to the bot are kept by Telegram in your chat with the bot under Telegram’s own rules. Frazo receives the audio from Telegram only to transcribe it and does not store it.

  3. 5.3

    A recording made on the site stays in your browser on your device until the server confirms it has received it.

Back to contents

6. Who receives your data

  1. 6.1

    We share data only with providers the service cannot work without, and only as much as each needs for its task. They act as our processors or under their own terms of service. The full list is published as Subprocessors and data locations (frazo.app/en/docs/subprocessors).

  2. 6.2

    We do not sell your data, share it for advertising or show one author’s data to another. We may disclose data to public authorities when the law requires it and the request is lawful.

Recipients of data

Recipients of data
RecipientPurposeDataCountry
DeluxhostServer hosting: storage and running the serviceAll service data: database, recording files, archives, backupsthe Netherlands
Groq, Inc.Speech recognition: transcribing voice answersThe audio of an answer and a list of niche terms, without your email or nameUSA
OpenRouter, Inc., and the model providers it routes requests toLanguage models: interview questions, drafts, revisions, memory factsTranscripts, author profile, topics, drafts, memory facts, without your emailUSA and the countries where the model providers operate
ResendSending email: sign-in codes, notifications, service messagesEmail address and the text of the messageUSA, with email sent through Ireland
TelegramThe bot, sign-in with Telegram, notifications, posting to your channelTelegram ID, messages and voice notes in the chat with the bot, published textsUAE and the countries where Telegram runs its servers
YANDEX LLC, Yandex MetrikaVisit statistics and behaviour analysis on the landing page of the Russian version: Session Replay (Webvisor), click and scroll maps, form analyticsIP address, cookies, browser, device and pages viewed, a session recording: pointer movements, clicks, scrolling, input in form fields except passwordsRussia
Google LLC, Google AnalyticsVisit statistics for the English version of the site, only with consentIP address, cookies, browser, device and pages viewedUSA
Back to contents

7. International transfers

  1. 7.1

    Our server is in the Netherlands. Speech recognition and language model providers are in the USA, our email provider Resend is a US company that sends our email through a data centre in Ireland, and Telegram processes data on its own servers.

  2. 7.2

    When data leaves the European Economic Area or the UK for a country without an adequacy decision, we rely on the safeguards offered by the recipient, such as the European Commission’s Standard Contractual Clauses or the recipient’s certification under the EU-U.S. Data Privacy Framework. You can ask us for details of the safeguards at the email below.

  3. 7.3

    We send providers only what their task needs. Speech recognition receives the audio of an answer and a list of niche terms, without your email or name. Language models receive transcripts, your profile, topics and drafts, without your email. With every request to OpenRouter we require that it is routed only to model providers that do not store requests and do not use them for training.

Back to contents

8. How long we keep data

  1. 8.1

    You choose the retention period for recordings under Language, theme and data: 7, 30 or 90 days, 30 by default. It runs from the answer, and for a draft you have not accepted, from its last change. Clean-up runs once a day; if you shorten the period, everything older is deleted at the next clean-up and cannot be restored.

  2. 8.2

    When the period ends, we delete the audio, the transcript, the topic titles and angles you typed, and drafts you have not accepted. Accepted texts and accepted memory facts with their quote and place in the recording stay: they are your archive.

  3. 8.3

    Database backups are kept for up to 7 days, so deleted data leaves the backups within 7 days.

Retention periods

Retention periods
DataPeriodWhat happens then
Audio of recordings made on the siteYour retention period: 7, 30 or 90 daysThe file is deleted
Transcripts and typed answersYour retention periodThe text is deleted; the date and length remain
Drafts you have not acceptedYour retention period from the last changeDeleted with their versions
Audio of Telegram voice messagesNot stored by usStays in your Telegram chat until you delete it
Accepted and published texts, accepted memory facts, speech profile, profile and settingsAs long as your account existsDeleted with your account
Data archive you request7 daysThe file is deleted
Sign-in session30 days, renewed while you use it, no longer than 90 daysYou need to sign in again
Sign-in and confirmation codes10 minutesThe code expires
Consent logWhile your account exists, and after deletion for as long as we need to prove consent and its withdrawalThe date, document version, IP address and email pseudonym remain, without a link to the account
Database backupsUp to 7 daysThe backup is deleted
Server logsA limited timeOverwritten by newer entries
Back to contents

9. Deleting recordings and your account

  1. 9.1

    Delete recording removes the audio and transcript at once, together with the drafts of that recording you have not accepted. Accepted texts and memory facts stay.

  2. 9.2

    You can delete your account under Language, theme and data. You confirm with a code, and the account is deleted 48 hours later; until then you can cancel. We tell you the time on screen and by email or Telegram message.

  3. 9.3

    Before deleting, you can download an archive of your data: texts with versions, transcripts, memory, profile, settings, topics, publications and voice recordings while they are kept. The archive is kept for 7 days.

  4. 9.4

    After deletion we keep only the record of your consent and its withdrawal, a deletion record without a link to you, and service cost records without your identity. Data already received by providers is deleted under their rules. Posts in your own channels are not deleted: they are your publications.

Back to contents

10. Your rights

  1. 10.1

    You have the right to access your data, to have it corrected, to have it erased, to restrict or object to processing, to data portability, and to withdraw consent at any time without affecting earlier processing.

  2. 10.2

    You can export your data and delete your account yourself under Language, theme and data. For anything else, write to us; we answer within one month and may extend this by two further months for complex requests, telling you why.

  3. 10.3

    You have the right to lodge a complaint with a data protection supervisory authority, in particular in the country where you live or work.

  4. 10.4

    Requests: hello@frazo.app. To make sure a request is yours, we may ask you to write from the email linked to your account or to confirm it in the dashboard.

Back to contents

11. How we protect data

  1. 11.1

    The site and dashboard work only over HTTPS. There are no passwords: you sign in with a one-time code by email or with Telegram, and codes are stored only as fingerprints.

  2. 11.2

    Voice recordings open only for their author, through a link that is valid for 10 minutes and checked together with your session. Files have no public addresses.

  3. 11.3

    Server logs contain no recordings, transcripts or drafts. Only the operator has access to the server and the database.

  4. 11.4

    If a personal data breach is likely to put you at risk, we will notify the competent authority and tell you without undue delay.

Back to contents

12. Cookies

  1. 12.1

    The site uses cookies to remember your language, theme, sign-in and analytics choice. Google Analytics runs on the English site only with your consent. Details are in the Cookie policy (frazo.app/en/docs/cookies).

Back to contents

13. Age

  1. 13.1

    Frazo is for people aged 18 and over. If you are younger, please do not sign up. If we learn that an account belongs to a minor, we will delete it.

Back to contents

14. Changes to this policy

  1. 14.1

    We may update this policy, for example when providers or the place of storage change. Each version is published on this page with a new number and date.

  2. 14.2

    If a change affects what data we process, why, who receives it or where it goes, we will tell you in the dashboard or by email before it takes effect and ask for consent again where needed.

Back to contents

15. Operator and contacts

  1. 15.1

    Frazo is operated by the owner of the frazo.app website (the “Operator”). The Operator is the controller of your personal data and the provider of the service.

  2. 15.2

    Contact the Operator, including for questions about your data, requests and withdrawal of consent, at hello@frazo.app.

  3. 15.3

    To ask for details about the Operator or about how your data is processed, write to hello@frazo.app.

Back to contents