← Navrick

Privacy

Last updated 26 July 2026

This page describes what the product does today, in the same words the landing page uses. It is not a substitute for a lawyer-drafted policy or DPA — those replace this summary before enterprise sales.

What we capture

Navrick mixes your microphone and your system audio on your Windows device to produce a meeting transcript and grounded notes, and it transcribes your dictation. It can also read the text visible on your screen so that names and jargon are spelled correctly.

Audio

A .wav of the meeting stays on your disk. We do not upload it, and we do not delete it for you. Transcription is a cloud call to Groq Whisper; that audio is transient. The file on your machine is yours to keep or delete.

Those files live in the app's data folder so a failed transcription can be recovered. There is no automatic retention today, and an explicit retention setting is planned. We do not claim "audio never leaves your computer" — transcription is a cloud service, and saying otherwise would be false.

Transcripts, notes and memory

Transcripts, meeting notes, dictation history and the facts Navrick remembers live in an encrypted local database on your device and nowhere else. They are not uploaded to Navrick. Transcript text is sent to an AI provider when a meeting ends, to produce the notes and the commitments; it is not retained there on our behalf.

An optional cloud mirror — so that a lost or reset machine is not a lost history — is in development and not shipped. When it ships it will be a visible setting, it will cover transcripts and notes but never audio, and this page will describe the retention that actually runs rather than the retention we intend.

Screen text

Screen reading happens entirely on your device using the Windows OCR engine. The image never leaves your machine and is never stored. A short list of rare terms may accompany a transcription request so that names come out right the first time.

The agent

Work that touches files or code is delegated to Claude Code, running under your own subscription and your own identity, in a sandbox. It cannot send, commit or publish anything without your explicit approval — that is a structural property of how it is wired, not a setting you can leave off.

Connectors

Notion, Slack, Teams and calendar are not built. No third-party workspace is connected, and nothing is pushed anywhere. When connectors ship, this page will list them before they are switched on.

Account

Sign-in happens in your browser through our auth provider; the desktop app never sees your password or code. Your account holds an email address and workspace membership — no meeting content. Your sign-in token and the key to your local database are stored in the Windows Credential Manager.

Subprocessors

Contact

Questions about data handling: emin@navrick.ai. One person reads that inbox, and it is the person who wrote the code.

See also Terms of use. The engineering contract behind this page is ARCHITECTURE.md §7 and §11a in the product repository.