How yospina-ctrl/jarvis-son-saldarriaga-ospina works
Jarvis-Son is a prototype of a lightweight, local Android voice assistant. Its intended workflow is listening, speech-to-text, response generation, and spoken output. The only sampled runtime source is a C++ orchestrator that simulates these states and prints status messages; it does not yet invoke capture, Whisper, an SLM, or Piper. The README confirms the product goal, but implementation details for the listed scripts, documentation, and model manifest were not sampled, so their runtime relationships are omitted.
Main entry points: main.cpp Cycle orchestrator
Components 5
Assistant control
- Cycle orchestrator
src/main.cpp
Voice workflow
- Audio capture
src/main.cpp - Speech transcription
src/main.cpp - Speech output
src/main.cpp
Response generation
- Response generation
src/main.cpp
Connections 4
- Cycle orchestrator → Audio capture: intended stage
src/main.cpp - Cycle orchestrator → Speech transcription: intended stage
src/main.cpp - Cycle orchestrator → Response generation: intended stage
src/main.cpp - Cycle orchestrator → Speech output: intended stage
src/main.cpp