Privacy and cookies

Last updated: 4 November 2026. Български

This explains what personal data auwo glas (glas.auwo.ai, the “Service”) processes, why, how long we keep it and your rights under Regulation (EU) 2016/679 (GDPR) and the Bulgarian Personal Data Protection Act.

1. Controller

Vaklin Stoynovski (natural person), Burgas, Bulgaria, trading under the auwo brand. Questions and requests: audit@auwo.ai. No data protection officer has been appointed; none is required for this kind and scale of processing.

2. Data we process

• Voice recordings you record or upload, and the text of them. Voice and words may contain personal data (customer names, phone numbers, addresses). If you speak about other people, you decide whether you may share that with the Service.
• Results: the table, note, text prepared for reading, and the generated MP3 file.
• Delivery address (optional): the email you want finished documents sent to, and a confirmation code.
• Account: email and name (and picture with Google sign-in), via Clerk.
• Usage record: how many seconds of audio you used in a month, for the 20-minute limit.
• Technical data: IP address, browser type, server logs and errors. We use no advertising or analytics cookies or pixels.

3. Purposes and legal bases

• To provide the Service: recording → document, text → voice (performance of a contract, Art. 6(1)(b)).
• To enforce the usage limit, protect the Service from abuse and fix faults (legitimate interest, Art. 6(1)(f)).
• To meet legal duties, including labelling of synthetic audio (Regulation (EU) 2024/1689, Art. 50).

4. Retention

• Recordings, texts, results and generated files: 30 days from creation, then deleted automatically (nightly). You can delete them earlier, one by one or with “Delete all mine” in the app.
• The monthly seconds record is kept while you have an account (it contains no content).
• The account is kept while you use it; to delete it write to audit@auwo.ai.

5. Who else sees the data (processors)

• fal.ai (USA) runs the speech-recognition and voice models; the recording and the text to be read pass through it. It runs third-party models (recognition: ElevenLabs Scribe; voice: Google Gemini TTS).
• Vercel (USA/EU): hosting, Vercel Blob (private audio storage, Frankfurt) and AI Gateway, through which the transcript text (no audio) goes to an Anthropic language model to be arranged into a table or note.
• Railway (Amsterdam, EU): the database with results and the usage record.
• Clerk (USA): sign-in and accounts.
• Microsoft 365 (EU): sends the mail, only if you switch on “Send by email”: a message with the finished document and the transcript text to the address you confirmed with a code. Messages are not kept in a Sent folder.

Transfers outside the EEA rely on the providers’ EU Standard Contractual Clauses or equivalent safeguards. We do not sell data or use it to train models; providers must not do so by contract.

6. Cookies

We use only strictly necessary cookies — Clerk’s, to keep you signed in. They need no consent, so there is no banner. We set no advertising, tracking or analytics cookies.

7. Your rights

You have the right of access, rectification, erasure, restriction, portability and objection, and to withdraw consent. You can delete content yourself in the app. For anything else write to audit@auwo.ai; we answer within one month. You may complain to the Bulgarian Commission for Personal Data Protection (cpdp.bg).

8. Synthetic voices

The voices are synthetic, not clones of real people. Every generated file is labelled — in its file name, inside the MP3 file (ID3 tag) and on the page where you receive it. Do not use it to suggest that a specific real person is speaking.