How gamkers/project_jarvis works
JARVIS supports web chat and desktop voice interaction, with the Flask API and standalone assistant both entering the core Jarvis logic. That logic handles agent-driven requests, task and reminder operations, conversation context, and semantic search; persistent data is stored in spreadsheets and vector files. README documents Gemini, Groq, and ESP32 Marauder capabilities. The source excerpts show API-to-assistant calls and some internal behavior, but not all agent-tool wiring or background-service details, so those relationships are omitted where unsupported.
Main entry points: jarvis_ai.py Agent tools · index.html Web UI · app.py Flask API · conversations.xlsx Conversation sheet
Components 20
User interfaces
- Web UI
static/index.html - Flask API
app.py - Voice mode
jarvis_ai.py
Assistant runtime
- Jarvis assistant
jarvis_ai.py - Agent tools
jarvis_ai.py
Assistant capabilities
- Conversation handling
jarvis_ai.py - Task management
jarvis_ai.py - Reminder management
jarvis_ai.py - User context
jarvis_ai.py - Vector search
jarvis_ai.py - Marauder control
jarvis_ai.py
Persistent data
- Conversation sheet
jarvis_data/conversations.xlsx - Task sheet
jarvis_data/tasks.xlsx - Reminder sheet
jarvis_data/reminders.xlsx - Vector data
jarvis_data/vector_db.pkl - FAISS index
jarvis_data/faiss_index.idx
External integrations
- Google Gemini
- Groq Llama
- ESP32 Marauder
Other components
- User
Connections 21
- User → Web UI: enters request
README.md - Web UI → Flask API: sends requests
README.md - Flask API → Jarvis assistant: calls assistant
app.py - User → Voice mode: speaks to
README.md - Voice mode → Jarvis assistant: runs assistant
README.md - Jarvis assistant → Agent tools: invokes agent
jarvis_ai.py - Jarvis assistant → Conversation handling: processes input
README.md - Agent tools → Task management: uses tools
README.md - Agent tools → Reminder management: uses tools
README.md - Agent tools → Vector search: uses tools
README.md - Conversation handling → User context: updates context
jarvis_ai.py - Conversation handling → Vector data: stores embeddings
README.md - Vector search → FAISS index: searches index
README.md - Task management → Task sheet: persists tasks
README.md - Reminder management → Reminder sheet: persists reminders
README.md - Conversation handling → Conversation sheet: stores history
README.md - Conversation handling → Google Gemini: generates responses
README.md - Agent tools → Groq Llama: uses model
README.md - Agent tools → Marauder control: offers security tools
README.md - Marauder control → ESP32 Marauder: controls device
README.md - Jarvis assistant → Web UI: returns response
README.md