AI-powered desktop dictation for Windows. Speak naturally — Stero Flow transcribes your speech with Groq Whisper, optionally polishes it through an OpenAI-compatible LLM, then types it into whatever app you're using.
GitHub link: https://github.com/navaneethdubbaka/Stero-Flow
For Users
Download
Get the latest release from GitHub Releases:
FileDescriptionSteroFlowSetup.exeOne-click installer (recommended)SteroFlow-portable.zipUnzip and run — no install required
If you built locally, artifacts are in dist/release/.
System requirements
- Windows 10 or 11 (64-bit)
- Microphone
- Internet connection
- Groq API key (free tier available) for transcription
- Optional: OpenAI-compatible LLM API key for Smart/Coding modes
- WebView2 Runtime (pre-installed on Windows 11; installer handles it on Windows 10)
Install (3 steps)
- Download and run
SteroFlowSetup.exe - If SmartScreen warns about an unknown publisher, click More info → Run anyway
- Launch Stero Flow from the Start Menu — it appears in the system tray
First-run setup
- The onboarding wizard opens automatically
- Paste your Groq API key and click Save & test
- Optionally configure an LLM key for polish modes (or skip)
- Press Start dictating
Using dictation
- Click into any text field (Notepad, VS Code, Chrome, Word, etc.)
- Press
Ctrl+Spaceto start, speak, thenCtrl+Spaceagain to stop - Text is transcribed and typed into the focused app after processing
ShortcutActionCtrl+SpaceStart / stop dictationCtrl+Shift+,Open settingsCtrl+Shift+MCycle dictation modeCtrl+Shift+PCommand palette
Portable version
- Extract
SteroFlow-portable.zipto any folder - Run
stero-flow.exe - No admin rights or installer needed
Troubleshooting
IssueFixSmartScreen warningClick More info → Run anyway (unsigned build)No microphoneSettings → Audio → select device; check Windows mic permissionsAPI key rejectedVerify Groq key at console.groq.com; for polish modes check LLM key in Settings → AIWebView2 missingRe-run installer or install WebView2 RuntimeText not appearingCheck Settings → Output is set to Type; some admin apps block keystrokesApp won't startSteroFlowSetup.exe is the installer — after install, launch from Start Menu. Check the system tray (^). See %APPDATA%\SteroFlow\logs\startup.log
See docs/USER_GUIDE.md for the full guide.
For Developers
GitHub link: https://github.com/navaneethdubbaka/Stero-Flow
Prerequisites
- Node.js 20+
- Rust stable with MSVC toolchain (
rustup default stable-msvc) - Visual Studio Build Tools — "Desktop development with C++" workload
- WebView2 Runtime
Clone and run
git clone https://github.com/steroflow/stero-flow.git cd stero-flow npm install npm run tauri dev
Build release installer
npm run build:release
On Windows, if MSVC link.exe is not on your PATH or you want Rust caches on E::
npm run build:release:win
This script sets CARGO_HOME=E:\.cargo, RUSTUP_HOME=E:\.rustup, CARGO_TARGET_DIR=E:\Stero Flow\src-tauri\target and loads the Visual Studio build environment automatically.
ArtifactPathSize (v1.0.0)Installerdist/release/SteroFlowSetup.exe~3.4 MBPortable ZIPdist/release/SteroFlow-portable.zip~4.7 MBRaw binarysrc-tauri/target/release/stero-flow.exe~8 MB
First release build compiles all Rust dependencies and may take 10–15 minutes.
Project structure
src/ React frontend (settings, overlay, onboarding) src-tauri/ Rust backend src/audio/ Microphone capture + VAD src/providers/ Groq Whisper transcription + OpenAI-compatible LLM src/secrets.rs API key storage (keyring) src/hotkeys/ Global shortcut definitions src/modes/ Dictation mode engine src/typing/ Text output (keyboard + clipboard) src/state.rs App orchestrator scripts/ Release packaging docs/ User guide, test matrix
Contributing
See CONTRIBUTING.md. Run npm run lint and cargo clippy before opening a PR.
Tech stack
- Tauri 2 — native shell via system WebView2
- Rust — audio, hotkeys, Gemini Live, typing engine
- React + TypeScript — settings and overlay UI
- Gemini Live API — real-time speech understanding
License
MIT — see LICENSE.