Tauri v2 + Rust Core

Speak your code.
Write at lightspeed.

A completely private, blazing fast desktop dictation engine.

System Architecture

SOLID Rust Architecture

STAGE 1 // INPUT

OS Boundary

Global low-level OS hooks via rdev. Emits filtered hotkey-pressed/hotkey-released intent events to Tauri for instant capture.

hotkey::HotkeyListenermod.rs

STAGE 2 // CORE

State Machine

A pure state machine serializing the lifecycle. Real side effects (like appending to the TranscriptionQueue) are decoupled into dedicated orchestrators.

orchestrator::TranscriptionCoordinatorcoordinator.rs

STAGE 3 // SENSORS

Audio Subsystem

Low-latency CPAL capture using safe lock recovery (lock_or_recover). Performs VAD and WAV encoding cleanly upon stop().

audio::streamstream.rs

STAGE 4 // CLOUD

AI Inference

Multipart async upload to Groq LPU (or custom OpenAI endpoints via api_url_override) parsed into a strictly-typed Result.

transcription::TranscriptionClientmod.rs

STAGE 5 // ACTUATOR

Output Engine

Surfacing back to OS. Simulates keystrokes using the abstract PlatformTyper trait, falling back to pasting if typing fails.

output::OutputHandlermod.rs

Why Voxis

Privacy, speed, and full control

Private by default

Audio goes straight to your chosen endpoint. No middleman, no telemetry.

Blazing fast

Groq LPU inference returns text in milliseconds. Dictation with no lag.

Local dictionary

Your own replacement dictionary and learning stay on-device — no cloud.

Any endpoint

Groq by default, or any OpenAI-compatible API via api_url_override.

Themeable

The overlay is fully customizable — themes edit live without a rebuild.

Cross-platform

Windows, macOS, Linux — one Tauri + Rust engine.

Voxis in action

An overlay on top of any app

Recording overlay

Transcription history

Settings & dictionary

Ready to write at lightspeed?

Free and open source.

Download Voxis Star on GitHub

Frequently asked

Where does my audio go?

Straight to your chosen transcription endpoint (Groq by default). Voxis does not proxy or store your audio on third-party servers.

Can I use my own server?

Yes. Any OpenAI-compatible Whisper endpoint via api_url_override in settings.

How do I set the hotkey?

In the app settings — record any combo; hold to start capture, release to finish.

Which OSes are supported?

Windows, macOS, and Linux — one Tauri v2 + Rust engine.

Speak your code. Write at lightspeed.