Hey, I'm Avi

I'm a 29-year-old guy who's really into coding, especially JavaScript. I love building handy little tools.
Find me here:

Stuff I've made


AnySTT Privacy Policy

Last updated:

I'm Avi - avi12 - and I write AnySTT on my own. It extends Windows 11's Voice Typing (Win + H) so it can dictate in languages Windows ships no speech model for. This is everything it does with your data.

What leaves your PC

While a Voice Typing session you opened is on screen, AnySTT streams the microphone audio Voice Typing is already capturing to the recognizer you picked in the Win + H gear menu, together with the BCP-47 tag of your active keyboard language so the recognizer knows what it is listening for.

Microsoft Copilot (default)
Goes to copilot.microsoft.com, signed in as your own Microsoft account through Windows' own sign-in.
Google Speech Recognition
Goes to speechs3proto2-pa.googleapis.com, signed in as nobody - no account, no sign-in.

What Microsoft and Google do with that audio is covered by their own privacy statements, not mine: Microsoft and Google.

Nothing goes anywhere else. I run no servers, so there is nowhere for your data to reach me. I get no audio, no transcripts, no telemetry, no crash reports, no analytics. I sell nothing, show no ads, and there is no account for you to create. Past the recognizer you chose, there is no third party in the path at all. Audio is streamed only while the Voice Typing panel is open; closing it ends the stream.

What stays on your PC

A Copilot access token, if you use Copilot
In %LocalAppData%\any-stt\, so you sign in once rather than every session. Deleted at uninstall.
Your settings
Two on/off values under HKEY_CURRENT_USER\Software\any-stt, and the recognizer you picked, stored with Voice Typing's own settings.
Language tokens
One per language you enabled, under HKLM\SOFTWARE\Microsoft\Speech_OneCore\UXLanguages\Tokens. This is the only thing AnySTT needs administrator for. Removed at uninstall.
A diagnostic log
%TEMP%\any_stt.log and %TEMP%\any_stt_watcher.log. It records what the app did, and can contain text you dictated. It never leaves your PC, I never see it, and you can delete it at any time.

Recording

AnySTT does not record your audio to disk. The one exception is a diagnostic you have to turn on yourself, by creating the file %TEMP%\any_stt_mic_dump.txt: while that file exists, the audio of each session is written beside it as .pcm so you can inspect what the recognizer actually heard. Delete the flag file to stop, and delete the .pcm files to remove the recordings. Those files stay on your PC too - send me one only if you want to, for a bug report.

Microphone

AnySTT never opens the microphone on its own. It reads the audio of the Voice Typing session you started with Win + H, for as long as that session is open.

Children

AnySTT is not directed at children, and I collect nothing about anyone's identity - from any age.

Questions

Email me at contact@avi12.com, or open an issue at github.com/avi12/any-stt-feedback.

YouTube Like-Dislike Shortcut

Shift+Plus to like, Shift+Minus to dislike. Can't get any simpler.

WebExt Deploy

A tool for deploying WebExtensions to multiple stores.

Skillshare Player Control

Control the Skillshare player just like you would YouTube's.

Twitch Channel Points Bonus Collector

Clicks the Channel Points bonus for you the moment it pops up.

Team Fortress 2 server.cfg Generator

Whip up a server.cfg for your TF2 server!

Counter-Strike 2 server.cfg Generator

Whip up a server.cfg for your CS2 server!

Garry's Mod server.cfg Generator

Whip up a server.cfg for your Garry's Mod server!