A modular, sarcastic, voice-powered Python assistant for Windows. Features real-time voice recognition, persona-based interaction, web search, and pluggable LLM completion (via Groq API).
- Voice and CLI interaction (Google/Edge TTS)
- Persona-driven sarcasm & wit (configurable)
- Integrated Brave web search
- Wake word and passive/active modes
- Fully modular, easy to extend
- Packaged as EXE via PyInstaller, or run as Python scripts
- Install Python 3.10+ and create a virtualenv:
python -m venv gda-venv gda-venv\Scripts\activate pip install -r requirements.txt
- Copy
.env.example
to.env
and fill in your API keys. - Copy
config.yaml.example
toconfig.yaml
and tweak as desired. - Run with:
python main.py
(from your venv), orrun_cortex_assistant_final.bat
(Windows)- Or build with PyInstaller for a single-file EXE.
Run:
pyinstaller --noconfirm --onefile --add-data "config.yaml;." --add-data ".env;." --add-data "google_creds.json;." main.py
Distribute the generated EXE with your config and keys.
- CLI: Type questions/commands. Type
listen
for voice input. Typeexit
to quit. - Wake Mode: Set
mode: "wake"
inconfig.yaml
to enable always-on listening for the wake word (hey cortex
).
Never commit .env
or google_creds.json
! All sensitive info is in .gitignore
by default.
MIT
See ROADMAP.md
for upcoming features!