Voice Type Privacy Policy
Developer: VNCNT Labs · Effective date: October 16, 2026 · Version covered: 1.0 (free) · Contact: [email protected]
This policy covers the browser extension VNCNT Voice Type: Dictation & On-Device AI Polish for Google Chrome and Microsoft Edge. Voice Type lets you dictate into text fields and then tidy the text. VNCNT Labs never receives your audio, your words or anything else you do with the extension. One thing you should know up front: by default, the speech recognition itself is done by your browser's speech service, which is a cloud service run by Google (Chrome) or Microsoft (Edge). Section 2 explains this and how to keep speech on your device.
1. When it listens and what it can access
Voice Type listens only after you start it (toolbar icon, Alt+Shift+D, or its mic button) and only while its blue bar is on screen. It stops when you press Stop, say "stop listening", switch tabs, leave the page, after a period of silence (45 seconds by default), or when the free daily minutes are used. The toolbar badge shows REC while it is listening. The microphone permission is granted to the extension itself, once, on its welcome page. Websites do not get microphone access through Voice Type.
It has no access to websites by default. When you start it on a tab, your browser gives it temporary access to that tab ("activeTab") so it can show its bar and type into the field you selected. If you turn on the optional Mic button on every site setting, your browser asks you for permission to run on all sites, only so that the mic button appears without pressing the shortcut first. You can turn that off at any time. Voice Type does not read pages, forms, passwords, cookies or browsing history. It only reads the few characters before your cursor in the field you dictate into, to get spacing and capital letters right, and the text you dictated when you ask to polish or replace it.
2. Speech recognition: browser service or on-device
- Browser speech service (default when on-device isn't installed). Voice Type uses your browser's built-in Web Speech API. In Google Chrome, this sends your audio to Google's speech servers, together with the browser language setting (Google states that cookies are not sent), and returns text. In Microsoft Edge, audio is processed by Microsoft's online speech service. That processing is governed by Google's or Microsoft's privacy policies, not ours. The extension shows a Browser speech tag while this engine is in use.
- On-device speech (when available). Newer versions of Chrome can download a speech language pack and recognise speech on your computer. Voice Type uses it automatically when the pack is installed (Settings → Dictation → Install on-device speech), and then shows an On-device tag. In that mode, audio does not leave your device. You can require on-device speech in Settings, in which case dictation will not start without it.
Either way, Voice Type itself never records, stores or uploads audio. The microphone is released as soon as dictation stops.
3. AI polish runs on your device
When you press Polish, your dictated text is processed by the AI model built into your browser (Chrome's Prompt API with Gemini Nano, or the Writing Assistance APIs where offered) on your computer. If that model is not available, built-in rules in the extension do the clean-up. Both run locally, and polished text is never sent to us or any AI service. If you click Enable on-device AI, your browser downloads its own model from its maker. That download is between your browser and Google or Microsoft.
4. Stored on your device only
- Settings: language, engine and polish preferences, shortcuts behaviour, mic button options.
- Usage counter: minutes dictated today, to apply the free plan's daily limit.
- History (optional, on by default): your last 10 dictations, so you can copy them again. You can turn this off or clear it in Settings → History.
- Session state: which tab and frame a running dictation types into. It is held in temporary browser session storage and cleared when the browser closes.
This data stays in the extension's local browser storage and is not synced to us. Uninstalling the extension removes it.
5. Pro (not yet available)
Version 1.0 is free only. It has no purchase or license features and makes no network requests of its own. The only network traffic is the browser's own speech service (section 2) and, if you click it, the browser's own model download (sections 2 and 3). When Pro launches, it will be sold through Lemon Squeezy, and activating it will send your license key and an instance name to Lemon Squeezy to check it. We will update this policy before that version is released.
6. Data sharing and selling
We do not collect, sell or transfer user data. We do not use any data for advertising, creditworthiness or lending, or for anything unrelated to dictation and text clean-up.
7. Chrome Web Store Limited Use
The use of information received from browser extension APIs follows the Chrome Web Store User Data Policy, including the Limited Use requirements.
8. Children
The extension is not directed to children under 13, and we do not knowingly collect information from children.
9. Security
The extension uses the "activeTab", "scripting", "offscreen" and "storage" permissions, plus optional access to all sites only if you turn that setting on. It runs no remote code: all scripts and fonts are packaged. Its pages use a strict content security policy, and its in-page interface is isolated from the page.
10. Trademarks
Voice Type is an independent product of VNCNT Labs. Google, Chrome and Gemini are trademarks of Google LLC. Microsoft and Edge are trademarks of Microsoft Corporation. They are mentioned only to describe which browser engines the extension uses.
11. Changes
If we change what the extension handles, we will update this page and the effective date before releasing that version.
12. Contact
VNCNT Labs · [email protected] · vncntlabs.com