Run Services Manually on Windows#
Use this guide when you want to run services individually for debugging or
custom development. For normal usage, prefer start_ata.ps1.
Scope#
This is a Windows-native workflow.
Prerequisite#
Clone and enter the application directory as shown in Get Started, then run:
.\setup_windows.ps1
setup_windows.ps1 prepares required submodules, virtual environments, and dependencies.
Service Startup Order#
Start each service in a dedicated PowerShell terminal.
1) metrics-collector (Port 9000)#
cd metrics_collector\windows
.\metrics_collector.ps1
Health check:
curl http://127.0.0.1:9000/health
2) text-to-speech (Port 8011)#
cd edge-ai-libraries\microservices\text-to-speech
.\venv\Scripts\python.exe main.py
Health check:
curl http://127.0.0.1:8011/health
3) audio-analyzer (Port 8010)#
cd edge-ai-libraries\microservices\audio-analyzer
.\venv\Scripts\python.exe main.py
Health check:
curl http://127.0.0.1:8010/health
4) rag-service (Port 8020)#
cd voice-enabled-interactions\smart-kiosk-assistant\rag-service
.\venv\Scripts\python.exe main.py
Health check:
curl http://127.0.0.1:8020/health
5) kiosk-core (Port 8012)#
cd voice-enabled-interactions\smart-kiosk-assistant
.\venv\Scripts\python.exe main.py
Health check:
curl http://127.0.0.1:8012/health
6) ai-teaching-assistant ui proxy server (Port 7860)#
cd .
voice-enabled-interactions\smart-kiosk-assistant\venv\Scripts\python.exe ata_ui_server.py
Health check:
curl http://127.0.0.1:7860/healthz
Open:
http://127.0.0.1:7860
Build React UI (if needed)#
If ata_ui_server.py reports that React assets are missing:
cd assistant-react-ui
npm install
npm run build
Then restart ata_ui_server.py.
Verify End-to-End#
Upload one or more documents in the Knowledge Base panel.
Click mic, ask a question.
Confirm transcript, answer text, and spoken response.
Stop Services#
Press Ctrl+C in each terminal.
Or use:
.\stop_ata.ps1
Notes#
The launcher flow (
start_ata.ps1) is the supported default.Manual mode is primarily for debugging service-level behavior.
Response audio clips are written under
generated_audio/in the kiosk-core area.