Legal
Privacy Policy
Effective July 19, 2026
Vocalis is built so that privacy is enforced by architecture, not by promise. Every decision to persist voice data passes through a single policy check in our codebase, classified by sensitivity. This page explains what that means in practice, in plain language.
What we never store
- Raw audio. Your voice is streamed for transcription and discarded. Audio buffers live in memory for the duration of a dictation session and are dropped when it ends.
- Screenshots. If a device sends a screenshot reference for context, it is stripped at ingest. It is never written to any store, log, or audit record.
- Sensitive context in audit logs. Admin actions are audited, but page content and transcript text never enter audit detail.
Data classes and how each is handled
Class A — Account data
Email, name, organization membership, settings, custom dictionary, snippets, styles.
Stored while your account exists. Needed to run the product and sync your devices.
Class B — Usage metadata
Feature usage, device platform, session counts, error diagnostics.
Stored in aggregate to operate and improve the Service. Never includes transcript text.
Class C — Session metadata
App name and type you dictated into, detected language, duration, word count.
Stored only when your settings allow it. Privacy Mode turns this off.
Class D — Dictation content
Final transcript text and surrounding context (page text, conversation snippets).
Stored only when both you and your organization's policy allow retention — and encrypted (AES-256-GCM) at the column level when it is. Privacy Mode guarantees it is never stored.
Privacy Mode
Turn on Privacy Mode and nothing from your dictations is persisted — no transcripts, no session metadata. Transcription happens in memory and the text lands at your cursor; that's the whole data path. Organization administrators can enforce Privacy Mode for an entire workspace, and that enforcement cannot be overridden by individual members.
Processors we rely on
We use a small set of subprocessors to run the Service, each receiving only what its job requires: speech-to-text and language-model providers process audio and transcript text transiently to produce your output; our database, cache, authentication, and hosting providers store and serve the data classes described above. We do not sell your data, and no subprocessor is permitted to use your content to train models without your consent.
Your rights
You can export your data or delete your account at any time from account settings; deletion removes Class A, C, and D data associated with you. Organization administrators can export or delete workspace data. Depending on where you live, you may have additional rights (access, correction, portability, objection) — write to us and we will honor them.
Changes and contact
If we change this policy in a way that matters, we will tell you before it takes effect. Questions or requests: sales@techsci.cloud. See also the Terms of Service.