Privacy policy

1. Summary

This privacy policy describes how Cao Hung Nguyen, as the operator of Rintori, processes personal data when you use the website, the application, the recording and import features, the AI features, payment functions or any other part of the service.

Rintori turns recordings you already have into a memory you can search and talk to. That means the most private material a person owns passes through it, so this page names every service that receives your content, what each one receives, and what it is used for.

• Your original audio is never overwritten. Transcripts, corrections and AI output are stored beside it.
• Speech-to-text and AI analysis are performed by external providers. Some of them are outside the EU. They are named individually below.
• We never sell personal data, and we do not use your recordings for advertising or to train third-party models on our instruction.
• The website itself sets no analytics or advertising cookies and loads no third-party fonts, scripts or embeds.

German and European data protection law comes first here. The GDPR, the BDSG and — for cookies, local storage and comparable technologies — § 25 TDDDG are the governing rules.

2. Controller

The controller within the meaning of Art. 4 (7) GDPR is:

Email: Email address: hung, at, theaiwhisperer, dot, dehung@theaiwhisperer.de

Because the controller is established in Germany, no separate representative under Art. 27 GDPR is required.

3. What we process

• Account data: email address, authentication data, Google sign-in data where used, language, theme, interface preferences and account status.
• Content you provide: audio recordings and imported audio files, transcripts, your corrections, typed notes, journal images, and files you attach to a recording as context.
• Data derived from that content: summaries, suggested tasks and calendar entries, speaker assignments, sentiment of the text, notes and Spark readings, and search indexes (vector embeddings).
• Voice profiles: if you enrol a speaker, a mathematical voice template. This is biometric data within the meaning of Art. 9 GDPR and is processed only on your explicit consent.
• Connected accounts: if you connect Google or Plaud, the tokens needed for the connection and the items you select for import.
• Usage data: feature events, processing stages, allowance and reservation events, and technical security events.
• Automatically collected technical data: IP address, browser and device characteristics, operating system, language settings, referring URLs, timestamps, log data and error reports.
• Billing data: plan, invoice and Stripe payment events. We do not process full card details ourselves.
• Optional location and weather context: only if you switch it on. It is off by default.

We do not deliberately ask for special categories of personal data under Art. 9 GDPR, but they can appear in a recording — a conversation about health, beliefs or a person's private life. Upload such material only where you are entitled to.

4. Purposes and legal bases

• Performance of a contract, Art. 6 (1) (b) GDPR: account, sign-in, upload and import, transcription, search, conversations with Tori, storage, allowances, payment and support.
• Explicit consent, Art. 9 (2) (a) in conjunction with Art. 6 (1) (a) GDPR: speaker recognition and voice profiles. You can withdraw it at any time, after which the template is deleted.
• Consent, Art. 6 (1) (a) GDPR: optional processing we ask about separately beforehand — microphone access in the browser, Google OAuth, connecting Plaud, place and weather context, and AI rewording of a spoken commitment into a task title.
• Legitimate interests, Art. 6 (1) (f) GDPR: security, abuse prevention, error analysis, keeping the service stable, and asserting or defending legal claims.
• Legal obligations, Art. 6 (1) (c) GDPR: tax and commercial retention, payment records, and statutory information, reporting and dispute-resolution duties.

Providing content is voluntary. Without an email address and a password you cannot have an account; without audio the service has nothing to work with. There is no automated decision-making producing legal effects within the meaning of Art. 22 GDPR, and no profiling for advertising.

5. AI processing and content

Rintori uses external AI providers to transcribe your audio, to produce titles, summaries, decisions, actions and insight from the transcript, to make your memories searchable by meaning, and to answer questions about them. Which providers, and what each receives, is set out in the next section.

• Audio is sent to the speech-to-text providers only. It is never sent to the AI provider, to the embedding provider, or to any other recipient.
• AI output can be wrong, incomplete or out of date. Generated content is labelled as generated, and an extractive excerpt is never presented as an AI summary.
• Generated output never replaces your evidence. Your original audio and the original transcript are kept intact beside it.
• You must not upload other people's personal data without a legal basis or their consent. For content you bring in yourself you remain responsible for its lawful source and use.

6. Service providers and recipients

These act on our instruction as processors under Art. 28 GDPR, or as independent controllers where stated. Each is listed with what it actually receives.

• Vercel Inc. (USA / EU region) — hosting, serverless functions, delivery and technical logs for the website and the application front end.
• IONOS SE (Germany) — the server that runs the application programming interface and stores your encrypted audio.
• Supabase Inc. (database region eu-central-1, Frankfurt am Main) — authentication, Postgres database, row-level security and the search index.
• Deepgram Inc. (USA; EU regional endpoint) — provides the product option called Transcription Pro and receives audio when it is selected. Rintori sends those requests only to Deepgram's EU regional endpoint and opts them out of model improvement. Deepgram documents that content on this path is processed and stored within the EU.
• Mistral AI SAS (France; EU regional endpoint) — provides the product option called Transcription Standard using Voxtral Mini Transcribe 2 and receives audio when it is selected. Mistral Small 4 is Rintori's default analysis model and receives the timestamped transcript to produce titles, summaries, decisions, actions and insight; individual passages together with your question when you ask something; and single sentences when you ask for a task title to be reworded. Mistral also turns transcript passages, your questions, Tori chat messages, the titles, start times and calendar names of events from the calendars you tick, and the file name and a short opening extract of an attached Google Drive document into search embeddings; event descriptions and guest lists are discarded before indexing, and a document's full text is never indexed. Regional inference is not the same as zero data retention, which is a separate account and contractual control.
• Modulate Inc. (USA) — legacy speech processor for recordings already queued before the Mistral migration. It is not offered for new recordings.
• OpenAI Ireland Ltd. / OpenAI, L.L.C. and DeepSeek (People's Republic of China) — optional analysis providers only when explicitly configured as the primary or fallback. In that case they receive the same bounded transcript material described for the analysis model above.
• Google Ireland Ltd. (Vertex AI, EU region europe-west4) — Live Talk: your voice, the bounded conversation context and tool results while a Live Talk session runs. Until the switch to Mistral, Google's Gemini API also produced the search embeddings described above; content sent to it before then was processed under Google's terms and cannot be recalled by us.
• Google Ireland Ltd. (Sign-in, Calendar, Drive) — only if you connect your Google account, and only for the scopes you approve. Our use of information from Google APIs follows the Google API Services User Data Policy, including the Limited Use requirements.
• Plaud — only if you connect a Plaud account, to import the recordings you select. Plaud processes this data on its own infrastructure, outside Rintori's EU processing.
• Brave Software, Inc. (USA) — only if you switch web search on (Pro) and confirm the disclosure. Receives the question as typed (up to 300 characters) and, for questions about places near you, your usual area rounded to about 11 km. Never a recording, transcript, person or account identifier. Brave does not publish where its API processes requests, so this is outside Rintori's EU processing.
• Stripe Payments Europe Ltd. — payment processing, billing and webhook events for paid plans. Billing records contain no recording title, transcript, attachment, voice template or derived memory content.
• Open-Meteo (Germany) — only if you switch weather context on. Receives a coordinate rounded to 0.1° (roughly 11 km) with a point in time, and separately a place name you type. It never receives a transcript, audio or an exact position.
• Mailjet (Sinch, EU) — sends account email such as sign-up confirmation and password reset on our behalf. Receives your email address and the message.
• Strato AG (Germany) — Rintori's own mailboxes and outgoing mail from rintori.com, including messages you send us and our replies.

7. Transfers outside the EU

Some of the providers above process data outside the EU/EEA, so parts of your transcript text and your audio are processed in third countries.

• United States (Deepgram, legacy Modulate work, Vercel, Stripe's group, Brave when you switch web search on, OpenAI where explicitly configured): transfers are based on the EU standard contractual clauses under Art. 46 GDPR and, where the recipient participates, on the EU–US Data Privacy Framework under Art. 45 GDPR. Deepgram transcription content is routed to its EU regional endpoint; Deepgram remains a US provider and EU data residency is not a sovereign-cloud guarantee.
• People's Republic of China (DeepSeek, only when explicitly configured): there is no adequacy decision for China. Transfers are based on the standard contractual clauses; be aware that the level of protection may not correspond to that of the GDPR, in particular regarding access by public authorities and the enforcement of your rights.

The default analysis path is Mistral's EU regional endpoint. This does not mean that every Rintori data flow stays in the EU or that Mistral has zero retention. Write to us and we will tell you which provider and retention controls are active at the time of your request.

We provide details of the transfer mechanisms and data processing agreements in use on request, where they are not already documented by the provider.

8. Google sign-in

You can sign in with Google where that option is shown. Google then supplies the profile data needed for sign-in and account creation — typically email address, name, profile picture and technical authentication data.

We use that data only for authentication, account management, security and the purposes described here. Google's own further processing is governed by Google's privacy information.

9. Cookies, local storage and tracking

We use cookies and local browser storage only where they are necessary for sign-in, security, language, theme, settings, sessions and allowances. The legal basis for accessing that information on your device is § 25 (2) TDDDG.

The public website sets no analytics, advertising or tracking technologies and embeds no third-party fonts, maps, videos or scripts. It stores one value — the language you chose — so a return visit opens in the same language.

Because nothing here requires consent under § 25 (1) TDDDG, there is no consent banner. If we ever add a technology that does require consent, we will ask for it before it runs and offer a way to withdraw it.

10. Retention

• Recordings, transcripts and derived content: until you delete them, or until you delete your account.
• Deleting a conversation immediately revokes playback and then removes its encrypted audio, its key, its transcript and its derived data, including the search index entries.
• Voice templates: until you withdraw consent or delete the person, whichever is first.
• Server, security and abuse logs: only as long as needed for the purpose they were collected for.
• Billing and tax records: the statutory retention periods under German commercial and tax law, separated from your memory data.

Where no contractual, legal or legitimate need remains, we delete or anonymise personal data. Backups may persist for a limited period for security reasons and are not used productively before they are deleted.

11. Security

• Transport is encrypted with TLS; stored audio is encrypted with a per-recording key.
• Provider keys stay on the server and are never delivered to the browser.
• Voice templates, embeddings, provider error bodies and your content are kept out of logs by design.
• The installed app's browser cache holds the application frame only — never a response containing your content.

12. Your rights

• Access to the personal data we hold about you (Art. 15 GDPR).
• Rectification of inaccurate data (Art. 16 GDPR).
• Erasure (Art. 17 GDPR) — you can also delete recordings and your whole account inside the product.
• Restriction of processing (Art. 18 GDPR).
• Data portability (Art. 20 GDPR).
• Objection to processing based on legitimate interests (Art. 21 GDPR).
• Withdrawal of any consent at any time, with effect for the future (Art. 7 (3) GDPR).

To exercise a right, write to Email address: hung, at, theaiwhisperer, dot, dehung@theaiwhisperer.de.

We may ask for proof of identity where that is necessary to protect your account, and we handle requests in line with applicable data protection law.

You also have the right to lodge a complaint with a supervisory authority (Art. 77 GDPR). The authority competent for us is the State Commissioner for Data Protection and Freedom of Information of North Rhine-Westphalia, Kavalleriestraße 2–4, 40213 Düsseldorf, Germany (https://www.ldi.nrw.de). A list of every EU authority is available at https://ec.europa.eu/justice/data-protection/bodies/authorities. You may also contact the authority where you live.

13. Recording other people

When you record or upload a conversation, other people's voices are in it. For that material you are the controller and we act on your instruction. Recording other people may require their consent, and the confidentiality of the spoken word is protected by criminal law in Germany (§ 201 StGB). Know the law where you are, and tell the people you record.

If you are in a recording someone else uploaded and want to exercise your rights, write to Email address: hung, at, theaiwhisperer, dot, dehung@theaiwhisperer.de — we will forward the request to the account holder responsible for it and support them in answering it.

14. Minors

Rintori is not aimed at children. Under our terms of use you must be at least 18 years old.

If we learn that we have processed a child's personal data without the required consent, we deactivate the account and delete the data unless a legal obligation prevents it.

15. Data protection officer

No data protection officer has been appointed; one is not required under Art. 37 GDPR and § 38 BDSG. Data protection enquiries go to the controller named above.

16. Changes

We update this page when the service, the providers we use or the legal position changes — in particular when a processor is added or the configured AI provider changes. The current version is permanently available at /legal/privacy, and the date at the top shows when it was last revised. Where a change requires your consent, we ask for it before the change takes effect.