How ujjwalagrawal3950/voice-assistance- works
Virtual KAM is a seller-facing commerce intelligence and AI account-management application. Sellers use the React workspace for dashboards, product and recommendation views, and text or voice conversations; the backend supplies deterministic business analysis and specialist-guided answers, while recommendations can be reviewed and acted on. Voice runs through a separate WebSocket service for transcription and speech synthesis, with recommendation actions relayed to the backend. The README documents the major capabilities; several API and domain modules were not sampled, so their internal call relationships are not inferred.
Main entry points: service.py Agent Coordinator · api REST API · main.py Voice WebSocket · client.ts REST Client
Components 25
Seller Workspace
- Workspace Routes
frontend/src/App.tsx - Chat Workspace
frontend/src/components/kam/KAMWorkspace.tsx - Voice Controls
frontend/src/hooks/useVoice.ts - Analytics Dashboard
frontend/src/components/dashboard/Dashboard.tsx - Recommendations UI
frontend/src/components/recommendations/RecommendationList.tsx - REST Client
frontend/src/api/client.ts
Commerce Backend
- REST API
backend/app/api - Commerce Analytics
backend/app/services/intelligence.py - KPI Analytics
backend/app/services/kpi.py - Forecasting
backend/app/services/forecasting.py - Recommendations
backend/app/api/recommendations.py - Commerce Database
backend/app/database.py
KAM Agent
- Agent Coordinator
backend/app/agent/service.py - Input Guardrails
backend/app/agent/guardrails.py - Domain Specialists
backend/app/agent/specialists.py - Business Tools
backend/app/agent/tools.py - Context Manager
backend/app/memory/context_manager.py - Conversation Memory
backend/app/memory/store.py - Memory Extractor
backend/app/memory/extractor.py
Voice Interface
- Voice WebSocket
voice-service/main.py - Speech Engines
voice-service/voice_engines.py
Other components
- Seller
- Gemini API
- Whisper STT
- Edge / Piper TTS
Connections 30
- Seller → Workspace Routes: uses
README.md - Workspace Routes → Chat Workspace: routes to
frontend/src/App.tsx - Workspace Routes → Analytics Dashboard: routes to
frontend/src/App.tsx - Workspace Routes → Recommendations UI: routes to
frontend/src/App.tsx - Chat Workspace → REST Client: sends chat
- Voice Controls → Voice WebSocket: opens WebSocket
README.md - Voice Controls → REST Client: sends text chat
frontend/src/hooks/useVoice.ts - Analytics Dashboard → REST Client: fetches analytics
- Recommendations UI → REST Client: manages actions
- REST Client → REST API: HTTP requests
frontend/src/api/client.ts - REST API → Commerce Analytics: serves analysis
- REST API → Forecasting: serves forecasts
- REST API → Recommendations: serves decisions
- Commerce Analytics → Commerce Database: queries records
- KPI Analytics → Commerce Database: queries records
- Forecasting → Commerce Database: uses seller data
- Recommendations → Commerce Database: stores decisions
- REST API → Agent Coordinator: dispatches chat
- Agent Coordinator → Input Guardrails: checks requests
backend/app/agent/service.py - Agent Coordinator → Domain Specialists: builds briefs
backend/app/agent/service.py - Agent Coordinator → Business Tools: runs tools
backend/app/agent/service.py - Agent Coordinator → Context Manager: assembles context
backend/app/agent/service.py - Agent Coordinator → Gemini API: requests reasoning
README.md - Context Manager → Conversation Memory: retrieves and stores
backend/app/memory/context_manager.py - Context Manager → Memory Extractor: extracts facts
backend/app/memory/context_manager.py - Voice WebSocket → REST API: posts chat
voice-service/main.py - Voice WebSocket → Recommendations: reads and updates
- Voice WebSocket → Speech Engines: synthesizes speech
voice-service/main.py - Voice WebSocket → Whisper STT: transcribes audio
README.md - Speech Engines → Edge / Piper TTS: generates audio
voice-service/voice_engines.py