How patchipup/jarvis-mark-ii works
Project type and purpose
• This is a local, modular, voice-controlled AI assistant (Jarvis) with no web UI. It runs as a desktop/CLI app, uses speech recognition and TTS, parses embedded commands, and integrates with external services (OBS, Spotify).
File structure
• Top-level orchestrator:
• main.py – bootstraps the system
• setup.bat / start.bat – Windows entry points
• config.ini + config.py – centralized settings loader
• input_listener.py – “press-and-hold T” detector
• audio.py – audio capture and TTS playback
• chat.py – AI/chat integration (e.g. OpenAI/Gemini)
• commands.py – parser for embedded in-text commands
• command_modules/ – dispatchers for each action type:
– main_module.py (general commands)
– obs_module.py (OBS WebSocket)
– spotify_module.py (Spotify Web API)
– music_module.py (local music folder)
– macro_module.py (user-defined command sequences)
• localization.py + locales/ – loads .po/.mo translations
• personality.py + personalities/ – loads persona JSONs
• tasks.py – scheduled or background tasks
• updater.py – self-update functionality
• Sounds & music/ – static media assets
README for additional insights
• Confirms modular design, lists speech engines (vosk, speech_recognition, faster_whisper), OBS & Spotify integration, macros, personalities, locales.