A Windows desktop assistant built with PySide6.
Now with Live speech recognition and Push-to-Talk (configurable hotkey) — the English Vosk model is bundled with the installer for plug-and-play.
- Modern GUI (PySide6)
- Live speech recognition (auto-finalize after brief silence)
- Push-to-Talk (press & hold) + configurable hotkey (e.g., Space / Ctrl+Space)
- Plug & Play installer (Inno Setup) — EN Vosk model included
- Clean packaging (PyInstaller onedir)
- Modular command/intent system (open apps/folders/websites, system info)
- Download the latest Atlas-Setup-*.exe from the Releases page.
- Run the installer (installs to Program Files, requires admin).
- Launch Atlas from the Start menu or desktop.
Quick test
- Live ON → speak → short pause → ✅ final text
- Live OFF → hold the PTT button or your hotkey → speak → release → ✅ final
- Say “help” → Help dialog opens
The installer also sets
ATLAS_VOSK_MODEL_DIRto the bundled model:
C:\Program Files\Atlas\models\vosk\vosk-model-small-en-us-0.15
VS_Atlas/ -atlas/ # App code (GUI, logic, speech, services) │ atlas_gui/ │ speech/ │ stt/
- assets/ # icons, etc.
- models/ │ vosk/ │ vosk-model-small-en-us-0.15/ # bundled speech model (via installer)
- dist/ # PyInstaller output (ignored in git)
- build/ # PyInstaller build cache (ignored)
- installer.iss # Inno Setup script (plug & play)
- requirements.txt
- README.md
- Windows 10/11 (64-bit)
- Microphone
- (For building from source) Python 3.11+, Inno Setup 6+ (for installer)
This project is licensed under the MIT License.
See the LICENSE file for details.